150
XIV. Cerebral Ischemia: The Role of Thrombosis and
of Antithrombotic Therapy
Study Group on Antithrombotic Therapy
EDWARD GENTON,
M.D.,
CHAIRMAN, 1
H. J. M.
BARNETT,
M.D., 2
MICHAEL GENT, 4 AND JOHN C. HOAK,
WILLIAM
S.
FIELDS,
M.D., 3
M.D. 5
Downloaded from http://stroke.ahajournals.org/ by guest on June 18, 2017
SUMMARY A revival of interest in antithrombotic agents for the
treatment of ischemic cerebrovascular disease has resulted in the
widespread use of oral anticoagulants for the prophylaxis and
therapy of the ischemic variety of stroke, and has generated
enthusiasm for the use of platelet-suppressant agents such as aspirin,
dipyridamole, and sulfinpyrazone. In delineating the several clinical
types of focal ischemic disease and outlining the causes of cerebral
ischemia and infarction, the study group considers the problems of
data interpretation in the face of inconsistent terminology. The basic
mechanisms of hemostasis and thrombogenesis are concisely detailed.
Finally, the study group critically reviews extensive earlier reports of
clinical trials of anticoagulants, platelet function suppressants, and
thrombolytic agents, and reassesses according to present-day
statistical standards the significance of the results. The information
contained in this report should familiarize the reader with sufficient
data to permit him to utilize antithrombotic agents under a variety of
circumstances and to appreciate the contraindications and potential
dangers in their use.
THE SOCIOECONOMIC importance of stroke is well
documented. 13 The economic impact in the United States
can be assessed from the figures provided by Cooper and
Rice3 who placed the total cost of stroke to the nation in
1972 at 6.2 billion dollars. Their calculation had three components: (1) direct costs for hospital care, physicians' services, nursing home care; (2) morbidity costs representing
lost earnings and dollar value of unperformed housekeeping
services; and (3) mortality costs referring to lifetime earnings (at present value) lost by premature death. Thromboembolism, in a comprehensive epidemiological review,4 accounts for the majority of vascular stroke. Recently, the
results of a prospective population study carried out in
Framingham, Massachusetts, analyzed with respect to cerebrovascular disease,8 substantiate that review. Over a period
of 18 years during which 5,184 men and women from the
community were followed, 196 became symptomatic as a
result of cerebrovascular disease. Of these strokes, 78% were
considered to be due to thromboembolic phenomena
(cerebral infarction due to thrombosis 57%, "transient ischemic attacks" 6%, and cerebral embolism 15%); 12% were
caused by subarachnoid hemorrhage, 5% by intracerebral
hemorrhage and the remaining 5% were ascribed to miscellaneous causes.
term special care because of physical difficulties or will be
bedridden and/or need institutional care.
It has been estimated 1 that 40% of the approximately
395,000 persons in the USA who have a thromboembolic
stroke each year will die within the first 30 days after the
stroke, and of those who survive more than a month
(numbering about 237,000), one-half will require either long-
'Professor of Medicine and Associate Dean, McMaster University Health
Sciences Centre, 1200 Main Street West, Hamilton, Ontario, Canada;
'Chairman, Department of Clinical Neurological Sciences, University
Hospital, 339 Windermere Road, P. O. Box 5339-Postal Station A, London,
Ontario, Canada N6A 5A5; 'Chairman, Department of Neurology, University of Texas Health Science Center at Houston, St. Anthony Center, 6301
Almeda Road, Houston, Texas 77021; 'Professor and Chairman, Department
of Clinical Epidemiology and Biostatistics, McMaster University Health
Sciences Centre, 1200 Main Street West, Hamilton, Ontario, Canada;
'Professor of Medicine, Director, Division of Hematology-Oncology, University of Iowa College of Medicine, University of Iowa Hospitals, Iowa City,
Iowa 52242.
Requests for reprints should be directed to: Joint Committee for Stroke
Resources, Suite 810, 1725 K Street, N.W., Washington, D.C. 20006.
Etiology
Cerebral emboli follow a path from heart-to-artery or
artery-to-artery. Classically, the heart has always been considered the source of a "cerebral embolus." Ten percent to
15% of cerebral infarcts are caused by embolism from the
heart. 4 ' 6 Rarely a thrombus will originate in the pulmonary
veins or a systemic venous clot will be carried through an
atrial septal defect. Cardiac emboli vary in size but commonly are large.
Mitral stenosis gives rise to one variety of large cardiac
emboli, and when accompanied by atrial fibrillation, the incidence of thromboembolism is said to be increased.7
Fibrillation is not a prerequisite to embolization; regular
rhythm was present in 17 of 72 cases of mitral stenosis and
cerebral embolism in one series.8 Forty-seven patients with
rheumatic heart disease were studied by McDevitt.9 The
hearts of 37 of these were in auricular fibrillation, and the
rest were in sinus rhythm at the time of the stroke. One
study10 indicates the incidence of systemic embolism (60%
cerebral) in surgically untreated patients to approximate
9.6%. In another series,11 cerebral embolism occurred in 34
of 172 (20%) rheumatic heart disease cases. Below the age of
35 years, the incidence of embolism is stated to be under 5%,
but above this age it rises beyond 30%.12
A patient with a major cerebral embolus from the heart
has a poor prognosis, with one in three experiencing an immediate mortality.13 The first year or two after the initial
embolus carries a greater risk of recurrence than is present
thereafter. Various reports give details concerning this outlook.10- 14"18
One of the best contemporary studies on survival, and
therefore prognosis, after the first cerebral embolism secondary to heart disease, is from a group of 42 cases followed
without treatment for 20 years. 8 At the end of this period,
one patient survived; 13 died of recurrent embolism, 22 of
congestive heart failure, and six of pneumonia and other
causes.
In a study14 of 194 patients with systemic arterial em-
CEREBRAL ISCHEMIA/7o/>i/ Committee for Stroke
Downloaded from http://stroke.ahajournals.org/ by guest on June 18, 2017
bolism in rheumatic heart disease, 30% of the 79 who had
single embolic episodes at the time of the first examination
died immediately. Of the total number of patients, 130 had
one or more episodes of cerebral emboli, and 64 of the 130
(49%) died as a result of this complication. In an attempt to
trace the fate of persons who survived the first episode, a
follow-up study" was made of those in whom the initial attack had occurred prior to the year 1940. This provided a
follow-up of ten years or to death as a minimum, with 19
years as the maximal period. Of the 52 known deaths during
this period, 14 occurred in the first year. In a further series15
of 125 patients with systemic emboli from mitral stenosis,
66% of the emboli were cerebral, and during the nine and
one-half year follow-up, recurrences developed in 30% of the
patients (35% of these within the first month of the initial
event; a total of 58% within the first year). In another study10
of 754 patients afflicted with established rheumatic cardiovalvular disease, 36 instances of cerebral embolism occurred
in those with atrial fibrillation, and 27 in those with sinus
rhythm. Among those patients who had recurrent embolism,
40% of the attacks appeared within one month and a total of
66% within 12 months of the initial embolus. The incidence
of emboli in this large series was calculated to be 1.5% per
patient-year. In another series16 of 17 patients (a control
group), 22 further embolic episodes occurred after the first
one during an average follow-up of five years, and 13 took
place in the first two years.
Mural thrombi associated with myocardial infarction are
a significant cause of stroke-producing cardiac embolism.
The risk of cerebral embolization within six weeks of acute
myocardial infarction may be as high as 5%. 17 ' 18 During the
immediate postischemic period about one-third of the patients die. Unless there is recurrent myocardial infarction,
the risk of further embolization decreases rapidly after four
to six weeks.
More than 100 years ago, pathologists noted that
obliterative atheromata in the carotid arteries in the neck
were associated with the findings of cerebral infarction. The
introduction and widespread use of angiography as a diagnostic technique, however, started the chain of events that
culminated in the realization that thrombotic lesions in the
cervical portions of the aortocranial arteries are commonly
associated with cerebral infarction. More recently,
awareness has increased that the extracranial arteries are the
site of stenotic and ulcerative atheromatous changes in
which thrombosis in situ may develop or from which embolization into the more distal intracranial portions of the
arteries may result. Accordingly, usage of the term, cerebral embolism, has been expanded to include artery-toartery emboli (intraarterial source).19 Such emboli are of
two main varieties which may exist in combination: plateletfibrin emboli and atheromatous material.
1. Platelet-Fibrin Emboli
White bodies were observed coursing through or lodging in
the retinal arterioles during an attack of amaurosis fugax.20
Later, this material was studied and shown to contain
platelets and fibrin.21 To identify platelet-fibrin emboli in the
vertebral-basilar territory is a more difficult task, since the
terminus of the vertebral-basilar circulation is not accessible
to direct visual examination as is the retina. One can
Resources
151
reasonably postulate that similar mechanisms operate in
both major arterial circulations to the brain, but in the
vertebral-basilar branches, this theory remains to be established.
2. Atheromatous Material
"Bright plaques,"22 as the atheromatous fragments in the
retinal arterioles have been called, may be seen when there is
ulceration of a plaque in the carotid arteries. Commonly
there is no special cause of the ulceration but it may be seen
after trauma to the carotid arteries. Following vigorous
palpation of diseased carotid arteries, and occasionally after
angiography or operation, a careful scrutiny of the fundi is
mandatory since cholesterol emboli may be present in the
retinal arterioles either with or without a history of
amaurosis fugax. Irregular atheromata and ulcer formation
may be visualized in angiograms of the neck vessels of patients with transient ischemic attacks (TIA).
Ulcerated lesions are located most commonly on the
posterior aspect of the lumen of the distal portions of the
common carotid and proximal segments of the internal
carotid arteries, but they can be identified laterally as well,
particularly in oblique views. Ulcers are observed much less
commonly at a distance from the cervical bifurcation, in the
carotid siphon, or in the vertebral arteries.
Statements have been made that the development of
occlusion in a carotid artery on the side appropriate to the
clinical signs and symptoms will lead to cessation of TIA.
Although this is generally the case, exceptions are well
documented that, subsequent to carotid occlusion, patients
have experienced ischemic episodes in the territory distal to
the obstructed artery.23 Several explanations are possible.
Hemodynamic factors may cause the blood pressure to drop
sufficiently to deny blood to the area of the brain now being
served by collateral circulation. Alternatively, the soft, friable tail of the occluding thrombus may break off and
obliterate a distal vessel that previously had been functioning well through collateral supply. The possibility also exists
that in the presence of internal carotid artery occlusion, emboli from a roughened or ulcerated lesion in the ipsilateral
common or external carotid artery, or in the contralateral
common or internal carotid artery may be carried distally
through various anastomotic communications. Finally, the
patent intracranial portion of a proximally occluded artery
may be the site of atheromatous disease, producing stenosis
and/or thromboembolism into the distal artery.
Incidence of Extracranial Versus Intracranial Lesions
Angiographic evaluation of the patients who were entered
into the Joint Study of Extracranial Arterial Occlusion24- "
demonstrated multiple atherosclerotic lesions and
emphasized the importance of the extracranial compared to
the intracranial vascular lesions (table 1).
On clinical grounds, even with the support provided by
angiographic and other examinations, one frequently encounters difficulty in determining whether individual patients have extracranial disease with embolization into the
intracranial arteries, intracranial disease with or without
embolization from the aorta and/or heart, or primary
obliterative disease occurring simultaneously in the extracranial and intracranial cerebral arteries. From the view-
152
STROKE
TABLE 1 Sites of Arterial Stenosis and Occlusion in Cerebrovascular Ischemia**
Site
Incidence of lesions*
Stenosis
Occlusion
Right Left
Right Left
Exlracranial arteries
Internal carotid bifurcation
Internal carotid (distal to
bifurcation)
Vertebral
Inlracranial arteries
Carotid siphon
Basilar
Middle cerebral
33.8
34.1
8.5
8.5
8.0
18.4
9.1
22.3
8.6
4.0
8.7
5.7
6.6
9.0
6.7
7.7
3.5
9.2
0.8
4.1
2.2
2.1
•Percent of lesions at designated sites in 4,748 patients subjected to
angiography.
Downloaded from http://stroke.ahajournals.org/ by guest on June 18, 2017
point of selecting appropriate therapy, these diagnostic
possibilities need to be examined as carefully as possible.
Despite painstaking study, however, frequently one is led to
the conclusion that, although reduced flow or obstruction is
present within a certain vascular territory, the exact etiology
of such ischemia is not identifiable with certainty.
Atherosclerosis is the most common cause of disease of
the aortocranial arteries and appears to generate most of the
clinically significant cerebral arterial occlusive events. Other
vascular lesions are rare, but include fibromuscular dysplasia, systemic lupus erythematosus, periarteritis nodosa,
giant-cell ("temporal") arteritis, allergic granulomatous
arteritis, Wegener's granulomatosis, and the vasculitis accompanying rheumatic fever, rheumatoid arthritis, or
scleroderma, as well as infections such as syphilis and
brucellosis.26 Other conditions, related more to abnormalities of clotting than to disease of the arterial wall, are rather
uncommon. Nevertheless they should be considered when
other evidence points to their presence. These include: (1)
hematologic diseases: primary thrombocytosis, leukemia,
sickle cell anemia, thrombotic thrombocytopenic purpura;
(2) disseminated intravascular clotting (DIC): heat stroke,
eclampsia, falciparum malaria, surgical shock, malignancy;
(3) hyperviscosity syndromes: polycythemia rubra vera,
Waldenstrom's and secondary forms of macroglobulinemia, cryoglobulinemia, multiple myeloma; and (4)
hereditary metabolic diseases: hereditary hyperlipidemia,
homocystinuria.
Venous Disease
Primary cerebral venous thrombosis is uncommon. For
the purpose of this review one should note that intracranial
venous occlusion occurs under all the circumstances in which
spontaneous venous thrombosis is known to arise,27 for example, in the veins of the pelvis and lower extremities. In the
presence of multiple etiologies accounting for systemic
venous thrombosis, the possibility of intracranial venous
thrombosis should always be kept in mind. Recent reports
indicate that the use of oral contraceptive medication is a
precipitating factor, thereby adding to the varieties of
thrombosis already known to be associated with hormonal
influences operating during pregnancy and the puerperium.
Clinical Varieties of Cerebral Ischemia
In the new Classification of Cerebrovascular Diseases19
the Clinical Stage (temporal profile) of ischemic stroke con-
TABLE 2
VOL 8, No
1, JANUARY-FEBRUARY
1977
Carotid Artery TIA: Presenting Symptoms in 133
Patients™
Symptom
Paresis (mono, hemi)
Paresthesia (mono, hemi)
Monocular visual
Paresthesia (facial)
Paresis (facial)
Dysphasia
Dysarthria
Headache
Lightheadedness
"Dizziness"
Convulsion (focal)
Convulsion (grand mal)
Binocular visual (hemianopia)
Visual hallucination
Dysphagia
Mental change
%Pta.
61
57
32
30
22
17
16
12
< 3 % each
tains a number of categories including transient ischemic attack (TIA), reversible ischemic neurologic deficit (RIND),
progressing stroke (stroke-in-evolution), and completed
stroke. Partial nonprogressing stroke is an added term for a
stroke that results in a minimal residual deficit.
Transient Ischemic Attacks (TIA)
Definition and
Symptomatology
These are episodes of temporary and focal cerebral dysfunction of vascular origin, rapid in onset (no symptoms to
maximal symptoms in less than five minutes and usually less
than a minute), which are variable in duration, commonly
lasting from 2 to 15 minutes but occasionally persisting as
long as a day (24 hours). The resolution or disappearance of
each episode is swift, ordinarily a few minutes at most. A
prolonged attack may take longer to clear. Usually, no
neurologic deficit remains after an attack. Transient ischemic events occur in the territory of the carotid and of the
vertebral-basilar arteries. Attempts to provide a clear understanding of the ischemic phenomena in each of these
territories and of their usual expression are important.
Otherwise, therapy may be misdirected and the critical appraisal of the efficacy of alternative treatment programs
may be erroneous. However, this distinction cannot invariably be made with certainty and some patients may, over
the course of a few weeks or months, manifest attacks indicative of ischemia in both the carotid and vertebral-basilar
territories. Table 2 is a tabulation of symptoms presented by
patients judged to be suffering from carotid territory TIA, as
recorded in one unpublished study.28 Table 3 enumerates the
symptoms exhibited by a group of patients from the same
unpublished study28 who were judged to have vertebralbasilar TIA. The main distinguishing feature, of course, is
the presence of unilateral symptoms in the carotid syndrome
(e.g., monocular blindness, hemiparesis) compared with the
bilateral symptoms in patients with vertebral-basilar syndrome. Furthermore, a number of isolated symptoms occurring without associated, more definitive manifestations cannot, of themselves, be attributed to TIA. These include
recurrent vertigo, diplopia, drop attacks, transient amnesia,
and episodes of unconsciousness. However, impaired con-
CEREBRAL ISCHEMIA//o//» Committee for Stroke Resources
TABLE 3
Verlebral-basilar TIA: Presenting Symptoms in 54
Patients™
Symptom
Downloaded from http://stroke.ahajournals.org/ by guest on June 18, 2017
Binocular visual
Vertigo
Pares thesia
Diplopia
Ataxia
Paresis
"Dizziness"
Headache
Nausea/vomiting
Dysarthria
Loss of consciousness
Visual hallucination
Tinnitus
Mental change
Dysphasia
"Drop attacks"
Drowsiness
Lightheadedness
Hearing loss
Hyperacusis
Dysphagia
Weakness (generalized)
153
TABLE 4 Hcmodynamic Causes of Diffuse Cerebral Ischemia
1. Orthostatic hypotension
a. Spontaneous
b. Iatrogenic
2. Carotid sinus hypersensitivity
3. Bradyarrhythmia — tachyarrhythmia
4. Valvular aortic stenosis
5. Angina with cardiac decompensation
6. Increased blood viscosity
7. "Steal" mechanism
8. Low perfusion states
9. Vasospasm
10. Vascular compression by musculoskeletal structures
%
57
50
40
38
33
33
20
18
14
14
14
7
5
5
3
3
< 3 % each
sciousness and drop attacks, as well as vertigo and nonspecific dizziness, when observed together and/or with other
brainstem symptoms, may present a picture sufficiently
definite to reinforce the diagnosis of TIA. For a full discussion of the definition of transient ischemic attacks, refer to
the Classification and Outline of Cerebrovascular Diseases
JJ 19(pp 594-596)
Pathogenesis
The pathogenesis of TIA falls into two major categories,
thromboembolic and hemodynamic.
Thromboembolic. Thrombi can form at the site of an ulcerated plaque or in an artery either at or beyond a stenotic
area. Many of these will not become sufficiently large to
constitute major emboli, but some will result in symptoms of
TIA or even a completed stroke, either by their transport to
intracranial branches or, in combination with an intruding
atheroma, by obliteration of the arterial lumen at the site of
intracranial lodgment. Rupture of a plaque or hemorrhage
into it can precipitate thrombus formation by causing excess
thromboplastin to be produced in the diseased arterial wall.
Embolism has been discussed in detail on pages 150 and 152.
This etiology is presently considered to be a most important
one for TIA because of the possibilities for instituting corrective medical or surgical therapy.
Hemodynamic. In every instance of TIA, an analysis
should be made to exclude all of the factors which could be
responsible for ischemia on a hemodynamic basis (table 4).
Many of these entities are accompanied by diminished cardiac output and are likely to be associated with diffuse disturbances of cerebral function, with impaired awareness,
and loss of consciousness. Should these predisposing factors
be operative in a patient simultaneously afflicted with
narrowed aortocranial arteries, an episode of focal ischemia
is possible, and indeed may be recognized as such. However,
in a review of a series of 290 patients29 who required a
pacemaker for control of cardiac dysrhythmia, only 1.4%
had evidenced TIA prior to therapy, in contrast to the 81%
with generalized neurologic symptoms, the most common of
which was syncope.
Increased viscosity of the circulating blood, as in
macroglobulinemia and polycythemia, is accompanied occasionally by secondary transient ischemic episodes. In the
steal mechanism, the radiologic identification of deviation of
flow from the brain toward an upper limb, the scalp, or the
face, as the result of subclavian, external carotid, or brachiocephalic artery occlusion is probably an uncommon event.
Most clinicians are agreed that for symptoms to result from
this altered flow alone is unusual, and that almost invariably
occlusive disease in the other extracranial arteries is
widespread.
The role of low perfusion has probably been overemphasized. In the human, a carotid artery must be very
severely narrowed before there is any change of pressure
gradient or of flow.30 Thus, diminished cerebral perfusion
resulting in focal cerebral symptoms appears to be uncommon. Furthermore, when multiple major arterial occlusions
are encountered and identified, they are frequently
associated with sufficiently well-developed collateral circulatory channels to prevent major infarction. Alternative
vessels appear to compensate satisfactorily for the potentially low perfusion until such time as either the collateral
arteries themselves become diseased, or cardiac function
becomes inadequate. TIA are not major features of stroke
symptomatology in patients known to have multiple major
arterial occlusions.
The possibility that vasospasm produces transient attacks
has not proved to be important in the etiology of TIA and
threatened stroke. Nevertheless, on rare occasions, the
sudden development of excessively high blood pressure, with
diastolic readings generally at or above 130 mm Hg, may be
associated with transient ocular and/or cerebral ischemic
episodes. In ophthalmoscopic examination during such a
vascular crisis, intensive retinal arteriolar narrowing is apparent. Management of the hypertension alone has been
observed to bring about cessation of symptoms and disappearance of the ocular changes.
The literature is replete with reports of TIA produced by
extravascular compression of the vertebral arteries by structures such as laterally placed osteophytes on the cervical
vertebrae. Tortuous, coiled, and kinked internal carotid
arteries have also been implicated. Despite a large number
of reports on the subject, extravascular compression has
proved to be a very uncommon cause of TIA.
The primary treatment of the hemodynamic group of con-
154
STROKE
ditions will, and indeed must, be directed at the responsible
circulatory mechanisms rather than at the thrombotic
process. Therefore, it is imperative to make this clinical distinction. Nevertheless, the impaired circulation occasioned
by disturbed hemodynamic phenomena may result secondarily in the initiation of thrombotic processes, especially
in the presence of diseased arteries. Indeed severe forms of
hemodynamic alteration can lead to massive cerebral infarction. White thrombi (largely platelets) appear in the
watershed pial arteries in patients dying from cardiac
failure, cardiac dysrhythmias, blood loss, and hypotension.31
Incidence, Natural History, and Prognosis
Downloaded from http://stroke.ahajournals.org/ by guest on June 18, 2017
Four outcomes are possible in patients with recurrent
T1A: (1) the attacks may cease spontaneously, (2) the attacks may continue, (3) cerebral infarction may occur, or (4)
the attacks may be associated with other major vascular
catastrophes (e.g., myocardial infarction), since the affected
population is at high risk of vascular disease.
Data are incomplete to delineate clearly the natural
history of TIA and the prognosis for patients with this disorder. A number of studies have been reported, but both the
clinical definitions employed and methods of compiling the
information lack standardization. Nevertheless, from
retrospective and a few prospective studies some useful information can be derived.
In retrospective studies of patients who have had a completed cerebral infarction, from 10%32 to as high as 79%33
were reported to have been afflicted with previous TIA. The
variability and range of the figures probably reflect the individual observers' interpretations of what has constituted a
preceding "attack," as well as the tenacity with which they
pursue the questioning with each patient.
Prospective studies may prove to be more accurate, but
again, the answers will be affected remarkably by the definition of TIA employed by the investigators. In a compilation
from several reported series34 with follow-ups ranging from
slightly less than one year to five years, the incidence of
cerebral infarcts following TIA varied, including 10% within
11 months, and 32% within 60 months. In a population
study,35 the clinical courses of 44 persons who had a transient ischemic episode were followed for periods averaging
27.4 months. Half of the number were given anticoagulants
and half were not. Of the 22 untreated patients, cerebral
thrombosis developed in seven (32%).
Studies are available to indicate that the first month
following the onset of TIA is the period of most serious risk.
A 15-year population study32 indicated that among the 72
patients with untreated TIA who had subsequent stroke, in
51% the stroke occurred in the first year, including 21% during the first month. A combined retrospective and prospective study36 indicated that if TIA were followed by a major
stroke, in 50% of the patients it occurred within one month
of the initial TIA.
A figure commonly cited suggests that 4% of the patients
will die each year after the first TIA, and most of those
surviving the immediate episode will die ultimately from
vascular causes, one-third of which will be from stroke and
two-thirds from other vascular disorders, particularly myocardial infarction.
In attempting to assess the overall significance of TIA, the
prognosis has been examined under three special sets of cir-
VOL 8, No
1, JANUARY-FEBRUARY
1977
cumstances. Firstly, in a five-year follow-up of 64 patients
with a history of TIA but with normal carotid angiograms,37 20% had further transient attacks, 13% had a
completed stroke, and 22% had fatal cardiovascular or cerebrovascular events, with heart disease occurring more frequently than stroke. Secondly, TIA confined to the retina
(amaurosis fugax) have been surveyed in a study38 involving
80 persons followed for periods of three months to ten years
or more. Five had a stroke and nine sustained permanent
visual loss during the period of follow-up, all in the territory
of the appropriate carotid artery. Thirdly, minimal information is available correlating the development of stroke with
frequency of TIA episodes. As reported in one series,39 76%
of those patients having transient focal ischemia in the
carotid territory, in contrast to 52% of those in whom the
vertebral-basilar system was involved, had stroke after only
one or two antecedent TIA. Further data are needed on this
important subject to obtain a reliable estimate of prognosis.
Reversible Ischemic Neurologic Deficit (RIND)
The distinction between RIND and TIA, as defined
previously, is one of degree.36 To discuss its nature, pathogenesis, and treatment in a separate way is probably redundant, because this condition may be only a temporal extension of TIA beyond the limits arbitrarily established for the
cessation of transient symptomatology.
Partial Nonprogressing Stroke
Partial nonprogressing stroke can be described as an
episode of localized cerebral ischemia, with a minor or
moderate residual neurologic deficit. Events such as these
may or may not be preceded by single or multiple TIA
without residual deficit. Many authors prefer to include this
condition in the "completed stroke" category.
Pathogenesis
The pathogenesis is presumed to be identical to that of
TIA and RIND. The difference depends upon the size of the
thrombus or embolus, the location and extent of the cerebral territory involved, and the condition of available
anastomotic channels at the time of interference with flow
through the primary vessels.
One particular kind of partial stroke which should be set
apart from the rest is the so-called "lacunar stroke." 40 This
type of cerebrovascular disease is associated with longstanding severe hypertension, and over a period of weeks or
months the patient has repeated minor ischemic episodes,
frequently with accumulating residual disability. The
etiology is still the subject of controversy because some
authors favor the concept of repeated small hemorrhages.
The severity is variable; some patients with episodes of
lacunar infarction have no symptoms whatsoever, while
others have many "small strokes."
Diagnosis and Prognosis
The criteria for diagnosis and prognosis for the patient
with a partial nonprogressing stroke and the principles of
therapy to be applied are the same as those for the patient
with TIA. In both instances appropriate management requires as exact and definitive a diagnosis as can be reached
with certainty.
CEREBRAL ISCHEMIA/7o/>if Committee for Stroke Resources
Progressing Stroke (Stroke-in-Evolution)
Downloaded from http://stroke.ahajournals.org/ by guest on June 18, 2017
In the past this clinical state has been imprecisely
described. A definition has been attempted and its acceptance urged so that there can be uniformity in reporting the
condition. 19 -" According to this definition, a progressing
stroke is one in which the neurologic deficit is still increasing in severity or distribution after the patient's admission to
observation, with evidence that this progression has been going on during the few minutes immediately preceding and up
to two or three hours previously. It is stated that only under
these circumstances is the pathophysiologic process a
progressing one, for which this diagnosis should be considered. Some would restrict the time interval of progression to a number of hours while others would consider it
reasonable to include those who worsen over a period of
days or even several weeks. The stuttering or stepwise stroke
progression would be included, to which the term "stroke-inevolution" has previously been ascribed. Jones and
Millikan42 have indicated that a progression-free period of
18 to 24 hours for the carotid system and up to 72 hours for
the vertebral-basilar system should ordinarily be sufficient to
remove the patient from this category and to label him as an
example of "completed stroke." Each author has utilized his
own definition of progressing stroke, thus precluding meaningful comparisons between reports of frequency, prognosis,
and results of therapeutic measures designed to affect this
condition, and leaving profound gaps in our knowledge.
Since some clinicians regard progressing stroke as an indication for the emergency use of heparin,43- ** a resolution of the
dilemma of terminology is an urgent matter.
Incidence
The incidence of progressing stroke has been variously
reported. In the Cooperative Study on Anticoagulant
Therapy,45 128 of 443 cases included in the overall study
were considered to be "progressive" or "thrombosis-inevolution." In this particular study, the definition of the
clinical entity was that of a progressive, stepwise lesion. In
one early report46 on carotid artery occlusion, it was stated
that of 35 cases, ten had a slowly progressive onset, and 22
an episodic (stuttering) course. In another report,43 122 of
277 patients admitted over a four-year period with a
diagnosis of cerebral thrombosis were classified as progressing stroke.
Pathogenesis
It is important to note that stroke-in-evolution does not
necessarily indicate thrombus-in-evolution. The variety of
pathogenetic mechanisms that can lead to either the slowly
progressive or the stepwise development of a stroke are summarized as follows:
1. A progressive thrombus extends from its site of origin
in a primary artery and obliterates collateral branches,
thereby interfering with anastomotic circulation and extending the area of insult.
2. At the site of maximal atherosclerotic involvement
with or without ulceration and/or stenosis, initially there is
insufficient thrombus to produce occlusion. Continuing accretions to this thrombotic nidus slowly obliterate the lumen
of the vessel and either gradually or intermittently add to the
area of brain ischemia.
155
3. Brain edema spreads in concentric fashion and
progressively reduces clinical function without extension of
the original area of infarction.
4. The general condition of the patient, including his cardiorespiratory function, altered water and electrolyte regulation, and/or acid-base balance, or the acquisition of
systemic infection, interferes with cerebral metabolism
sufficiently to increase the extent of the neurological dysfunction. In the event that cardiorespiratory function is
altered, the area of infarction actually may be extending.
The problems of pathogenesis are seen to be complex, and
thus decisions respecting therapy in this condition become
very difficult. Serial follow-up of these patients by means of
computed tomography is an encouraging area for research
in order to clarify our understanding of the pathological
alterations in stroke, in stroke progression, and in the
regression of brain lesions.
Course and Prognosis
Progressing stroke has a poor prognosis. Of 41 patients
entered into the National Cooperative Study and reported in
196147 as having "thrombosis-in-evolution," 30% worsened
in the two weeks after recognition of their condition, and
between two weeks and six months thereafter another 17%
deteriorated. In a British study" where the symptoms of
ischemic deficit had been progressing over a period of two
hours or more, a control series of 38 of the 76 cases was
followed for six months. At the time of follow-up, 11 had
recovered, 8 had improved, 12 were unimproved, and 7 had
died. In an analysis of five series,34 followed for periods
ranging from six months to 15 months, the percent of patients not treated by anticoagulants who had progressed to
deterioration or death ranged from 40% to 64%. The conclusion can be drawn that the recognition of a progressing
stroke, whatever its cause, carries an uncertain prognosis.
Completed Stroke
The term "completed stroke" is usually employed to
signify a focal neurologic disability that came on abruptly
and has become stabilized. Some authors34 indicate that 18
to 24 hours without progression if the lesion is in the carotid
system, and up to 72 hours if the territory of involvement is
the vertebral-basilar system, would suffice to place the patient in the stabilized category. It is also noteworthy that
only a gradation in severity differentiates a partial nonprogressing stroke from a completed stroke.
The variability in symptoms and in the degree of initial
deficit, as well as the variability in recovery of lost function,
are dependent on a number of factors, many of which are
similar to those which determine the "progression" of a
stroke. These include the adequacy of collateral circulation,
the state of cardiorespiratory function, and the presence or
absence of major systemic dysfunction that might critically
alter cerebral metabolism. Unilateral carotid occlusion and
even, on occasion, bilateral carotid occlusion can sometimes
be silent events. Angiographic studies performed on these
patients subsequently often show remarkable collateral circulation in many of the deficit-free individuals. In those with
a fixed and serious neurologic deficit, the channels of the
potential collateral circulation, although demonstrable, apparently did not function with sufficient rapidity to protect
the territory supplied by the occluded vessel.
156
STROKE
Pathogenesis
5
Downloaded from http://stroke.ahajournals.org/ by guest on June 18, 2017
The Framingham study indicated that 78% of vascular
strokes in that population were thromboembolic and that
thrombosis of the aortocranial arteries occurred between
three and four times as frequently as did strokes from cardiac embolism. Embolus from the pulmonary veins is an extreme rarity, as is paradoxical embolism. The thrombotic
process may be: (1) initiated and completed in the heart,
with fragments subsequently lodging in the great vessels or
their intracranial branches; (2) initiated, completed, and
fixed in the great vessels of the neck; (3) initiated in the neck
and carried to the intracranial branches; or (4) initiated and
completed in the intracranial branches.
Regardless of the source of the thrombus, thefinalresult
is an area of brain infarction whose extent will depend on the
size of the embolus, on the immediate availability and
utilization of collateral circulation, and on the state of
systemic oxygenation and circulation.
It is important to reemphasize that the majority of patients with completed stroke will have a history of one or
more previous TIA. Furthermore, for practical purposes,
one should recognize that a stroke may be regarded as
"completed" because of a stabilized neurologic deficit and
yet another stroke of abrupt or stepwise progression, or a
subsequent TIA may occur. An apparently simple extracranial carotid lesion may thus be responsible for major or
minor embolization repeated over a long or a short period of
time.
The preceding resume underscores the importance of
thrombosis as a prime factor in the pathogenesis of a variety
of cerebrovascular disorders. Clearly the intelligent management of threatened and developed cerebral ischemia requires
an understanding of thrombosis and of antithrombotic
therapy.
Hemostasis and Thrombogenesis
Since the time of Virchow, it has been appreciated that
alterations in the vessel wall, increased coagulability of the
blood, and vascular stasis operate to provoke a thrombotic
episode. In experimental thrombosis, usually two of the
above components of the triad are sufficient to provide a
suitable model for study of the process of thrombogenesis.
Hemostatic and Thrombotic Mechanisms
The hemostatic-thrombotic mechanism acts to prevent
loss of blood from the circulation by rapid repair of breaks
in vessel walls with a minimal compromise of blood flow.
Hemostasis and thrombosis involve interaction between: (1)
the blood vessel, (2) platelets, and (3) coagulation factors, to
form a fibrin-platelet patch of the defect, which is then subjected to modification by thefibrinolyticsystem prior to and
during tissue repair (fig. 1).
Blood Vessels
Vascular damage elicits responses of all components of
the hemostatic system. Vasoconstriction follows direct injury to the vessel and is associated with constriction of adjacent vessels. Although vasoconstriction is not required for
hemostasis, it is a critical reaction to prevent rapid ex-
VOL 8, No 1, JANUARY-FEBRUARY 1977
sanguination following the severance of large vessels. When
the endothelial barrier is removed or disrupted, platelets
adhere immediately to the exposed subendothelial connective tissue structures including basement membrane,
microfibrils, and collagen fibers. Vasoconstriction is
enhanced by the release of serotonin and epinephrine from
the aggregated platelets and by substances such as
fibrinopeptides formed during the generation of fibrin.
Platelets
In hemostasis, the platelets play a crucial role consisting
of: (1) initial arrest of bleeding by formation of a platelet
plug, (2) contribution of a phospholipid to the coagulation
process for stabilization of the hemostatic plug, and (3) support and maintenance of endothelial integrity by a
mechanism incompletely understood.
Platelets circulate as ovoid cytoplasmic disks, 2 to 4 n in
diameter, at a concentration of 250,000 ± 40,000 platelets
per microliter of blood, and have a life span of nine to ten
days. In normal persons, two-thirds of the total platelets are
in the general circulation while the remaining one-third is
held in the spleen as a pool of platelets that exchange freely
with those in the general circulation. The morphologic
features of the platelet are shown infigure2 (A and B). The
peripheral surface coat (plasma membrane) mediates the
contact reactions of adhesion and aggregation and also
forms an open membrane system that expands the reactive
surface into which plasma hemostatic factors can be absorbed.
When platelets come in contact with collagen, striking
changes occur. The platelets swell and irreversible fusion of
the accumulated platelet mass into a hemostatic plug takes
place.
Platelet consumption may occur as part of intravascular
coagulation or, independent of coagulation, as an effect of
contact with normal surfaces (e.g., prosthetic heart valves,
bacteria).49 Even with extensive vascular injury, such as that
found in vasculitis, direct consumption of platelets can occur
at sites of endothelial detachment without an appreciable
decrease in clotting factor activity. This mechanism may
operate also in patients with thrombotic thrombocytopenic
purpura, the hemolytic-uremic syndrome, and vasculitis induced by toxic, immunologic, or metabolic factors. Acute
arterial thromboembolism and recurrent venous thromboembolism also are associated with increased platelet consumption.
Vessel Wall Injury
Platelets
Adhesion to Collagen or
Subendothelial Tissues
Coagulation
Activation
(Extrinsic)
Platelet Aggregation (AOP Release)
(Intrinsic)
- Fibrin
Platelet Plug (or Thrombus)
A
Stable Platelet Plug (or Thrombus)
Fibrinogen
FIGURE 1.
Hemostatic plug formation.
CEREBRAL ISCHEMIA/Zo/n/ Committee for Stroke Resources
157
Downloaded from http://stroke.ahajournals.org/ by guest on June 18, 2017
--Mt
FIGURE 2. Transmission electron micrographs of a human platelet section in two planes. (A) X 36,000; (B) X 25,600. C = canaliculus,
DB = dense body, G = granule, Gly = glycogen; M = mitochondrion, Mt = microtubules.
Coagulation Factors
Blood clotting involves a series of enzymatic reactions
which converts circulating inactive coagulation proteins to
active forms culminating in the transformation of blood
from a liquid to an insoluble gel through the conversion of
fibrinogen into fibrin (fig. 3). Fibrin formation extends and
stabilizes the hemostatic plug in its evolution to a thrombus.
Fibrin formation can be effected by either of two major
overlapping mechanisms termed intrinsic and extrinsic. All
the coagulation factors required for the intrinsic system to
function are present in circulating blood; the extrinsic
system, by means of a lipoprotein released from damaged
cells ("tissue factor"), provides a bypass of the usual intrinsic activation pathway. Factor X plays a central role in
coagulation since it can be activated by either the intrinsic or
extrinsic pathway.
Fibrinolytic
System
The fibrinolytic system plays a major role in the
maintenance of blood fluidity, particularly in the small
vessels. Vascular endothelium contains a potent tissue activator which converts an inert plasma protein, plasminogen,
into the potent enzyme plasmin (fig. 4). Activation of the
fibrinolytic mechanism is linked inevitably to activation of
the hemostatic process, and fibrinolysis should be considered
an integral component of hemostasis. An inverse relationship has been demonstrated between vessel size and fibrinolytic activity. Thus the small vessels appear to be more resistant to occlusive thrombus formation than the large ones.
Protective Mechanisms Against Activated Coagulation
Several physiologic mechanisms are known to protect
against thrombogenesis: (1) An increased rate of blood flow
reduces the chance of focal fibrin formation by removing
procoagulant material before activation can be completed
and by diluting substances that may have undergone activation. (2) Activated clotting components are removed from
the blood rapidly by hepatic cells. Paniculate material, including fibrin, is taken up by reticuloendothelial cells, and
activated coagulation factors, especially Xa, are removed by
hepatocytes. (3) Thrombin and Xa are inactivated slowly in
vivo by the formation of complexes with a physiologic
plasma inhibitor. This inhibitor, referred to as antithrombin III or factor Xa inhibitor, also functions as a
heparin cofactor in a reaction involving ionic binding of
heparin. Of interest are reports of a high incidence of
recurrent venous thrombosis in families among whose
members are a large number having decreased antithrombin III activity.
158
STROKE
INTRINSIC
COAGULATION
EXTRINSIC
COAGULATION
Tissue Factor
Factor VII*
Factor
Factor
Factor
Factor
X I I - » - Xlla
XI - * - Xla
I X ' - * ~ IXa
VIII V
HEPARIN + ATI
Factor X * " — - ^
^-—Factor
Factor V
Phospholipid
Prothrombin*
Thrombin
Downloaded from http://stroke.ahajournals.org/ by guest on June 18, 2017
"I
Fibrinogen
Fibrin monomer
Ca
Fibrin
"Synthesis inhibited by
oral anticoagulants
(Warfarin)
Blood coagulation pathways.
Thrombosis and Blood Coagulation
A thrombus is defined as a solid mass or plug formed during life within the heart or vessels from constituents of the
blood (fig. 5). A blood clot is a blood mass solidified under
static conditions outside the body. Consideration of the
different types of thrombi is important from the standpoint
of their histologic structure and the sites and nature of the
flow conditions under which they develop. In a clot, all the
formed elements of the blood are distributed at random in
the fibrin meshwork. Platelets are found in the meshwork,
but do not form significant aggregates. In contrast, a thrombus consists of two types of material, usually referred to as
white thrombus and red thrombus. White thrombus consists
of closely aggregated platelets (figs. 6 and 7). Red thrombus
fills the interstices between the platelet aggregates and is
made up chiefly of red cells and fibrin. The type of thrombus
formed is related to the velocity of blood flow. White thrombi form at sites of more rapid flow (arteries, arterioles) while
red thrombi or mixed thrombi develop in sites where stasis is
likely to occur (veins, venules). The composition of the
thrombus also has implications for therapeutic approaches.
Fibrin formation is inhibited best by anticoagulants which
interfere with the formation and action of thrombin (e.g.,
heparin, warfarin), and therefore these agents may be appropriate for the management of venous thrombosis. On the
other hand, antiplatelet agents (e.g., aspirin, dipyridamole,
sulfinpyrazone) should be more effective against arterial
thrombi. The situation is highly unlikely to be as simple as
this, since thrombin takes part in the formation of arterial
platelet thrombi, and it has been shown that platelet adherence and aggregation participate in deep vein thrombosis.
Thrombosis can be related to atherosclerosis in at least
two possible ways. First, the "encrustation" or "thrombogenic" theory suggests that thrombosis is involved in the
pathogenesis of the atherosclerotic lesion. Second, whatever
the origin of atherosclerosis may be, thrombosis is clearly a
most important complication of the established lesion and is
responsible for a majority of the lethal and disabling
features of the disease.
Von Rokitansky60 was the first to suggest that the lesions
of atherosclerosis were caused by the deposition of material
from the blood onto the inner walls of affected arteries.
Interest in this hypothesis declined until, in 1946, Duguid51
put forward the theory in a revised form which has since
come to be known as the thrombogenic theory of atherosclerosis. Duguid, studying serial sections through stenosed
coronary arteries, found gradations between tissue that had
all the characteristics of an organizing thrombus and that
which had the appearance of a typical atherosclerotic
plaque. Despite the careful and comprehensive morbid
anatomical studies, acceptance of this theory has had to
await experimental proof that the end result of the organization of a mural thrombus is a lesion which has the detailed
features of an atherosclerotic plaque.
Several key points essential to the thrombogenic theory
have been proved by experiment. These include the demon-
STIMULUS
(Vessel injury, etc.)
COAGULATION
EXTRINSIC
PATHWAY
ACTIVATORS
lUrokinase. Streptokinase)
PLASMIN
(Active Enzyme)
+ Fibrin.
Fibrinogen
Other Proteins
INHIBITOR
Epsilon Aminocaproic
Acid
FIGURE 4.
I
FIBRIN
I
ADHESION
RELEASE
->- (ADP. PF3. etc.)
I
AGGREGATION
THROMBUS
SPLIT PRODUCTS
Fibrinolytic pathways.
PLATELETS
INTRINSIC
PATHWAY
THROMBIN
PLASMINDGEN (Inactive Form)
1977
Thrombosis and Atherosclerosis
Ca++
FIGURE 3.
1, JANUARY-FEBRUARY
X'
Activated Factor X (Xa)
[HEPARIN + AT^IUl—»>
VOL 8, No
FIGURE 5.
Thrombus formation.
CEREBRAL ISC H EMI A /yom/
Committee for Stroke
159
Resources
PI
RBC
Downloaded from http://stroke.ahajournals.org/ by guest on June 18, 2017
FIGURE 7. Transmission electron micrograph of a platelet thrombus in a small pulmonary blood vessel. PI = platelets; RBC = red
blood cell; VW = vessel wall; X 4.970.
FIGURE 6.
Scanning electron micrograph of platelet aggregates
stration that endothelium can grow over the surface of an
organizing mural thrombus so that the lesion appears to lie
within the intima rather than to be attached to its inner surface.52 The endothelium of large blood vessels is capable of
substantial regeneration and of behaving as the thrombogenic theory requires. Another serious question has been
answered concerning the presence of smooth muscle cells
that lie beneath the endothelium in atherosclerotic plaques.
Electron microscopic evidence has established that smooth
muscle cells are found in organized thrombi and, therefore,
could represent a source for those observed in evolving
atherosclerotic lesions. However, additional data to explain
the quantity and type of lipid found in advanced atherosclerotic lesions are required for support of the thrombogenic theory. The phagocytosis of platelets by monocytes
may contribute lipids to organized thrombi.53
It seems likely that the organization of mural thrombi
contributes to the enlargement of already existing severe
lesions. Whether the organization of mural thrombi can account for the total development of the atherosclerotic lesion
from beginning to end is debatable. Before significant
progress can be made in understanding this process, it is essential to learn more about the conditions under which
thrombi form spontaneously.
The atherosclerotic lesion is prone to complications arising from loss of the endothelial cover and the subsequent
development of thrombi. Atheromatous "gruel" causes
platelet aggregation and has a procoagulant effect, including that modulated by tissue thromboplastic activity. In
addition, platelets are capable of adhering to the subendothelial structures including collagen, microfibrils, and endothelial basement membrane. Although undoubtedly thrombosis complicates atherosclerosis, additional investigations
are necessary to elucidate: (1) the magnitude of and the
mechanisms involved in platelet adherence to intact en-
dothelium, (2) the thrombogenic complications of vascular
injury, and (3) the efficacy of agents used to prevent these
complications.
Diagnostic Tests
Despite the urgent need for tests to detect the individual in
whom thrombosis is impending or incipient, progress has
been slow in the development of procedures which are both
accurate and practical for large numbers of patients.
Procedures that prove useful for the diagnosis of peripheral
venous thrombosis may have little value for the patient with
an impending stroke or TIA.
While importance has been attached to the association in
various clinical and pathophysiologic states between excess
coagulant activity and a tendency to develop thrombosis, the
term "hypercoagulable state" is probably best reserved for
the circumstance in which activated coagulation factors are
present in the blood. These activated factors may not, of
themselves, precipitate massive thrombosis or even disseminated intravascular coagulation (DIC), but may
become thrombogenic under certain conditions, such as
vascular stasis or impairment of physiologic inhibitory
mechanisms. In addition, certain activated clotting factors
(thrombin or Xa) have the potential to cause aggregation or
adherence of platelets.
Of the blood tests available which may be useful in selecting stroke-prone patients, most are directed at measurements of platelet function, the presence of fibrinogen
derivatives as a result of in vivo thrombin activity, or
changes in fibrinolytic activity. The use of routine coagulation tests, such as the bleeding time, prothrombin time, clotting time, tests of platelet adhesiveness or platelet aggregation with aggregating agents, is rarely of help in predicting
thrombosis. Other approaches include the measurement of
5
'Cr-labeled platelet survival, detection of "circulating"
platelet aggregates,54 testing for "spontaneous" platelet
aggregation using platelet aggregometry, and the use of a
160
STROKE
screen filtration method for platelet aggregation.55 The latter
three tests have demonstrated abnormalities in patients with
cerebral ischemia and may be useful to identify patients who
are likely to respond to antiplatelet agents. 5458 However,
prospective studies remain to be performed with these
techniques to establish a cause-and-effect relationship
between either the findings of platelet aggregates associated
with TIA symptoms and signs or the disappearance of
clinical signs and symptoms with subsidence of platelet aggregates as a result of drug therapy.
Therapeutic Inhibitors of Thrombosis
Downloaded from http://stroke.ahajournals.org/ by guest on June 18, 2017
The oral anticoagulants, of which warfarin is an example,
exert their effects by inhibiting the synthesis of coagulation
factors II, VII, IX and X in the liver. Warfarin inhibition of
the action of vitamin K causes defective factor synthesis
resulting in the production of dysfunctional zymogens that
are antigenically identical with their corresponding clotting
factors but lack coagulant activity. There is a time lag
between the beginning of therapy with warfarin and attainment of the desired depression of clotting factor activity.
Likewise, restoration of coagulation activity to normal is
delayed when vitamin K.! is administered to overcome the
anticoagulant effect of warfarin.
Heparin, unlike warfarin, causes no deficiencies of
coagulation factors, but instead interferes with the coagulation process at several points as follows: it inhibits the activation of factor IX by XIa; it inhibits IXa, thrombin, and
Xa; and in the absence of thrombin, activation of factor
XIII is prevented. Heparin also strikingly enhances the
ability of antithrombin III to inhibit the action of activated
factor X (Xa).57
Considerable interest has developed in drugs which interfere with platelet function and thus might serve as effective antithrombotic agents, but the role of these preparations
in the management of cerebrovascular disease is still being
evaluated. Platelet aggregation is blocked by various pharmacologic agents, the most potent of which is prostaglandinEi.5B This effect is associated with an increase in cAMP
(cyclic adenosine monophosphate) concentration in the
platelet. Other drugs include aspirin, dipyridamole, sulfinpyrazone, and phenylbutazone.
Aspirin interferes with the production of certain intermediate metabolites in platelets, and this intracellular
change may be associated with changes in platelet function.58 The effect of aspirin on platelets occurs as a consequence of acetylation of the platelet membrane. Therefore,
the analgesics, sodium salicylate and acetaminophen
(Tylenol), do not interfere with platelet function. Other
platelet inhibitors include adenosine and its analogues,
sulfhydryl-binding agents, and numerous other drugs that
block glycolysis and oxidative phosphorylation. Antiplatelet
agents vary in the duration of their effects. In the case of
aspirin, this lasts for the remainder of the life of the exposed
platelet, whose total life span is eight to ten days. As a consequence of the theoretical implications, the use of antiplatelet agents in antithrombotic therapy has excited intense
interest. Whether currently available preparations will fulfill
the expectations of their proponents remains to be determined.
VOL 8, No
1, JANUARY-FEBRUARY
1977
Evaluation of Antithrombotic Therapy
An appreciation that thrombosis or thromboembolism is
etiologic in many cerebral ischemic events has encouraged
the employment of therapeutic approaches intended to influence the thrombotic process. These have included the use
of anticoagulant, platelet-suppressant, or thrombolytic
drugs, depending upon clinical circumstances and
therapeutic objectives. In the following pages, reported
studies for each of these categories are reviewed and an
attempt is made to determine from the data presently available the potential of the respective treatments for preventing or modifying the results of an ischemic stroke.
Design of Clinical Trials
A number of basic principles should be considered in the
design and execution of clinical trials. In general, there
should be a clear statement of the objectives of the study,
together with a detailed description of the type of patient under study, the actual maneuver to be evaluated, the outcome
variables of interest, and the methods of analysis. Specificity
should be sufficient to allow replication. Within this general
framework, the types of studies which have been carried out
include individual case reports, retrospective studies, and
prospective studies, both randomized and nonrandomized.
Genton et al.,59 in a critique of studies of plateletsuppressant therapy in clinical thrombotic disease, have
identified and discussed certain basic principles for clinical
trial design. Among these, randomization and sample size
are of particular relevance to this evaluation of antithrombotic therapy. Unless the patients are allocated
rigorously to their respective treatment according to some
prescribed randomized arrangement, there is the danger that
bias, either unconscious or deliberate, might influence which
treatment a particular patient receives. In the nonrandomized trials to assess anticoagulants, for example, it is
possible that patients too sick to receive anticoagulants are
left untreated and then are included as controls; or a patient
who experiences another cerebrovascular episode or dies
while under evaluation as one to receive anticoagulant
therapy might be listed as an untreated subject, and hence
introduce an obvious bias to the disadvantage of the control
group record. This also implies that studies should be
prospective, with controls which are concurrent in time and
place.
Reports of individual case histories or of accumulated experience with a series of patients in which the observed
therapeutic benefits are compared either with those noted
during the author's previous experience, or with those of a
concurrent group of patients who were ineligible to receive
that therapy, have the strong possibility of bias, and the findings from these studies can be regarded as only suggestive.
One encounters a particularly difficult problem in the
evaluation of antithrombotic therapy in cerebrovascular
ischemia because of the large number of patients required to
achieve statistical significance. Since the immediate death
rate from cerebral thrombosis is lower than in other types of
stroke, long-term follow-up and large-scale multicenter
trials are necessary. For example, the overall frequency of
stroke in patients with transient ischemic episodes is ap-
CEREBRAL ISCHEMIA/7o/w/ Committee for Stroke Resources
proximately 7% each year.34 Thus, even if a drug brought
about 50% reduction in the incidence of stroke, more than
1,000 patient-years would be required for a study to be
statistically meaningful! Even trials of secondary stroke
prevention require a study of comparable size, since the frequency of recurrent stroke is similar.80
Evaluation of Reports
Downloaded from http://stroke.ahajournals.org/ by guest on June 18, 2017
The following review deals in order with the studies relating to anticoagulant, platelet-suppressant, and thrombolytic
therapy, respectively. Within the first and largest category,
that relating to anticoagulants, attention is given separately
to TIA, progressing stroke, completed stroke, and cerebral
embolism. Studies are considered individually and include a
description and critique of the design, methodology, and
results. The accumulated evidence is summarized as to
evidence of possible therapeutic benefit derived by patients
within each of the four functional categories of stroke. The
studies are few from which to evaluate antiplatelet and
thrombolytic therapy and therefore a single summary is
provided for each of these two types of agents.
One should bear in mind that most of the studies of anticoagulants were carried out 15 years ago, and are here being
judged against standards and methodologies which have
been improved and refined in recent years. What is being
presented, therefore, is not a denigration of work done in the
past, but rather a reassessment, by present-day standards, of
the data which bear upon the efficacy of antithrombotic
therapy in cerebrovascular ischemia. The discussions to
follow are not intended to reflect upon the interest,
capabilities, or competence of the investigators, but to
provide information which will be of value in future clinical
research.
Anticoagulant Therapy
Rationale
In patients with cerebral ischemia, the objective of
therapy acutely is to preserve the integrity of the ischemic
tissue as far as possible, and to prevent progression or recurrence of the insult. When the genesis of an ischemic event is
thrombotic or thromboembolic, theoretically the objective is
to prevent further development and propagation of the
thrombus so as to minimize tissue damage and, in the long
term, to prevent recurrence of local thrombosis of cerebral
vessels and/or reembolization from a distant site.
The clinician has drugs available which have been demonstrated experimentally to arrest active thrombosis and to
decrease the incidence of recurrence. Heparin, by combining with a plasma globulin, forms a complex which rapidly
and reliably inhibits several activated clotting factors, including the thrombin presumably present at the site of active
thrombosis. By inhibiting thrombin activity, further fibrin
formation is prevented and the propagation of thrombosis
interrupted. Consideration of such treatment is rational for
patients suspected of having an active thrombotic state.
In other circumstances, such as after recovery from an
ischemic stroke, the therapeutic objective is prevention of
recurrent thrombosis or embolization of a cerebral vessel.
Experimentally, it has been established that by interfering
with the hepatic synthesis of vitamin K-dependent clotting
161
factors, oral anticoagulants (e.g., coumarin) can retard the
response to a moderate thrombotic stimulus and in that way
act as prophylactic agents.
Randomized Trials of Anticoagulants in Patients With TIA
The Veterans Administration Cooperative Study61 was a
multicenter randomized trial to evaluate the benefit of oral
anticoagulants in patients with TIA or with cerebral infarction. Most of the cases were entered into the study within
one month of onset of their last cerebral episode. Uniform
laboratory and clinical criteria were established for the inclusion and exclusion of patients — among the latter, those
with severe hypertension and blood in the spinal fluid.
Angiograms were not performed to determine the location
of arterial disease. Anticoagulation was monitored using the
one-stage prothrombin time method and in 80% of patients
good control was maintained. The study was not blind and
the treated patients were seen more frequently than the untreated, so that the interpretation of recurrent symptoms
was not free from bias, although this should be regarded as
less important, inasmuch as mortality and cerebral infarction were the principal end points. Over a two-year period,
1,409 patients in all were screened, of whom 155 were
accepted for entry into the study; their assignments were
randomized into treatment and nontreatment groups. The
two groups were comparable in general and all the patients
were male. No patient was observed for less than one month,
the average follow-up period being 12.8 months in the control and 9.3 months in the treated group.
The outcomes for the 37 patients with TIA are summarized in table 5. One can see that only one patient in the
treated group had a new ischemic attack, compared with
eight patients in the control group. The figures on stroke and
death are too few for sound inferences to be drawn. Figures
on bleeding were not given separately for the patients with
TIA and those with completed stroke. There were ten major
bleeding complications in the treated group, including four
fatal massive hemorrhages, compared with only three in the
control group, one of which resulted in death. Members of
the treated group experienced many more minor hemorrhages than did the controls.
Baker, Broward, Fang et al." reported a multicenter trial
carried out between 1958 and 1961 in seven clinics. Patients
were categorized as having: (1) TIA, (2) thrombosis-inevolution, (3) thrombosis — completed stroke, (4) "thorem"
(thrombosis or embolism), and (5) cerebral embolism. Inclusion and exclusion criteria were specifically defined, lumbar
punctures were carried out, but angiograms were not
utilized. Hypertensive patients were not excluded. Only persons whose last cerebrovascular episode had occurred within
the eight weeks before randomization were admitted to the
study. Those qualifying for inclusion were randomized into
either the anticoagulant or the control group, the latter
receiving placebo capsules. The comparability of the two
treatment groups was based on sex, race, age, blood pressure, territory of brain damage, a neurologic disability scale,
time since the qualifying neurologic episode, previous
stroke, and smoking. In the follow-up, both control and
anticoagulant patients returned at the same intervals for
evaluation of clinical status.
162
STROKE
VOL 8, No
1, JANUARY-FEBRUARY
1977
TABLE 5 Randomized Clinical Trials of Anticoagulants
Average
Study
Baker et al. (VA)«
Baker, Broward,
Fang et al.46
Pearce et al.62
Baker, Schwartz, Rose63
Carter". a
Baker, Broward,
Fang et al.46
Marshall, Shaw69
Baker et al. (VA)«
Baker, Broward,
Fang et al.46
Hill, Marshall, Shaw70
Downloaded from http://stroke.ahajournals.org/ by guest on June 18, 2017
71
Howard et al.
McDowell, McDevitt 72
Enger, Boyesen73
No. pts.
T* C*
period
(mo.)
C
T
1961 TIA
22
15
9
13
1
1962
1965
1966
1961
Stroke-in-evolution
24
17
30
38
20 18
20 11
30 38
6
38
21
11
41
6
10
10
—
1962 Stroke-in-evolution
1960 Completed stroke
61
26
1961 Completed stroke
56
67 12
25
2
6
62
9
1962 Completed stroke
1962 Completed stroke
Phase I
Phase II
1963 Completed stroke
1965 Completed stroke
1965 Completed stroke
72
60
71
66
15
92
51
71
65
15
99
49
Yea
of
report
Condition
TIA
TIA
TIA
*T = group receiving anticoagulants; C
tFirst month only.
TIA
C
T
Lethal
strokes
T
c
Results
All
Ischemic
deaths
Btrokes
T
C
C
T
Bleeding
M inor
Major
T
T
C
C
0
0
1
0
1
0
21
9
14
—
—
0
0
—
—
1
1
—
1
1
4
2
4
—
5
0
9
2
3
5
7
8
0
3
—
1
—
0
0
0
0
1
—
15
1^
6
13
— —
— —
— —
1
3
6
—
—
6
10
—
—
2
17
3
4
7
20
—
—
21
1
—
—
2
2
—
—
10
1
—
—
3
11
11
—
—
6
9
28
12
42
23
10
31
12
34
23
—
—
—
6
4
—
—
—
9
6
4
5
—
4
1
8
2t 10f
5
—
8 21
— —
— —
3
2 10
1
0
3
0
3
6
4
13
6
2
12
5
12
6
18
15
24
2
7
0
0
1
—
9
22
—
20
5
4
19
—
22
10
8
1
4
3
57
6
—
12
3
53
6
—
0
3
12
0
5
4
0
0
0
2
0
7
3
8
3
21
7
0
11
3
•• control group.
A total of 443 patients was studied. Of the 44 who had
TIA, 24 received anticoagulants and 20 were in the control
group. Recurrence of TIA was much more frequent in the
control group — 547 episodes compared with only 25 in the
treated group during average follow-up periods of 18 and 21
months, respectively. In the first month, ten control patients and two on treatment continued to have attacks. Two
of the 20 patients in the control group died, compared with
five of the 24 receiving treatment; two of the latter deaths
were due to intracerebral hemorrhage. Progression or recurrence of infarction occurred four times in the control group.
Three of these infarctions were moderate and one mild compared with the single severe case in the treated group. There
were 11 patients in the treated group who had bleeding
problems, three of them severe, as compared with one minor
bleeding complication in the control group.
Pearce et al.82 carried out a randomized double-blind trial
with 37 patients aged less than 70 years who had had one or
more attacks of transient cerebral dysfunction with no
neurologic deficit lasting more than 24 hours. Patients with
malignant hypertension or peptic ulceration were excluded.
The clinical examination included assessment of cerebrospinal fluid; carotid angiography was carried out in 29 of the
patients. None had significant narrowing of the extracranial
portions of the internal carotid arteries. Patients were randomly allocated to a high-dosage (50-mg tablets of phenindione) or an ineffective low-dosage (1-mg phenindione
tablets of identical appearance). Seventeen patients received
phenindione in effective dosage over an average period of
11.1 months; 20 control patients were followed for an
average of 10.6 months.
The two groups were comparable with respect to age and
sex, but a few more patients with vertebral-basilar artery insufficiency were included in the control group than in the
high-dosage group. Ten patients in the treated group (59%)
had further TIA, compared with nine in the controls (45%).
One patient in each group sustained a nonfatal completed
stroke; one control patient died from a stroke and two others
died from cardiac causes. None of the patients had major
hemorrhagic complications, and no instance of minor bleeding was definitely attributable to anticoagulant therapy. The
average period of follow-up was less than a year, so that the
study was a limited one, and therefore the benefits of treatment would have to be fairly spectacular in order to be established as statistically significant.
Baker, Schwartz and Rose63 reported an eight-year randomized study designed to evaluate long-term anticoagulant therapy for patients with TIA. Sixty of 204 individuals were selected after excluding those who were poor
risks for anticoagulation; these were the severely hypertensive persons, those critically ill, near death, and those of advanced age (more than 80). The patients were all male
veterans, their ages averaging 62 years. During the study,
the desired therapeutic range of prothrombin time (determined by the one-stage method) was achieved an average
79% of the time, and for most patients, for 90% of the time
or better. Each patient was examined neurologically once a
month during the first few months of follow-up, and then
every two months. Only ten of the 60 patients had arteriography. There was no report that a placebo was given to the
control patients and no evidence that the assessments were
made blindly.
The 60 patients were randomized into groups of equal
size. The average period of follow-up was 38 months for the
treated and 41 months for the control group; 10 and 14 patients respectively in the two groups had further TIA. There
were nine new cerebral infarctions. Of the four in the control
group, two recovered without neurologic disability and two
had only mild functional deficit at the end of the study,
whereas of the five in the treated group, two were functionally normal and three had severe residual deficits. One
should note that in the last three cases the infarctions occurred after anticoagulants had been stopped (for reasons
unspecified). It is not clear in the report how long after dis-
CEREBRAL ISCHEMIA/yo/nf Committee for Stroke Resources
Downloaded from http://stroke.ahajournals.org/ by guest on June 18, 2017
continuance of therapy the infarctions happened, or why the
anticoagulants were discontinued. If the time between
therapy withdrawal and infarction was short, or if the reason
for withdrawal of anticoagulants related in some way to the
clinical condition of the patient, then these therapeutic
failures should be charged against the treatment. In the long
run, this may constitute a bias against showing a favorable
treatment effect — it is judged, however, that this is
probably the preferable direction in which to err. The alternatives are to omit these cases entirely from the analysis or
to charge them against the controls, since the patients were
no longer being treated. This is likely to result in a bias
toward an apparent benefit of treatment.
Of the 14 deaths, nine were in the treated group and only
five in the control group. It is interesting that nine of the 14
deaths were the result of cardiac disease. A few minor
hemorrhagic complications arose, but only one patient (in
the control group) had a fatal cerebral hemorrhage. The
authors noted that the limitations of their study included the
small number of cases and the large number of variables in
patients with TIA which might influence prognosis.
Nonrandomiied
TIA
Trials of Anticoagulants in Patients With
Fisher64 reported on 29 patients with TIA who were
treated with dicumarol for periods varying from two months
to four years (average 14 months). TIA ceased in 28 of the
29 patients after effective anticoagulant therapy had been established; only one patient had a stroke, and that nearly
three years after beginning treatment. Following cessation
of treatment, 12 of 20 patients had recurrent episodes. When
anticoagulation was resumed, the attacks ceased. The
criteria for stopping and starting treatment were not stated.
In contrast, a group of 23 patients was studied for whom
anticoagulants were not prescribed for various reasons. Of
these, 12 had further TIA over an unspecified period of
follow-up and eight patients had a stroke (six of a major
type).
This study raises a number of questions. Patients were not
randomized to treatment and control groups, nor was a
rationale given for deciding whether or not a patient would
receive anticoagulants. Considerable bias is possible in that
one-third (mostly those admitted late in the study period)
were followed for some time before a decision was made to
place them on anticoagulants. The description of the patients was generally rather imprecise and none was excluded
on the basis of age, blood pressure, or location of the lesions.
As the author himself said, "The present data do not provide
incontrovertible evidence of the therapeutic efficacy of anticoagulants because satisfactory control patients have not
been studied."
Siekert, Whisnant, and Millikan65 from 1954 to 1958
observed 335 patients with TIA for whom anticoagulant
therapy was prescribed except when specifically contraindicated. Anticoagulants were administered to 175 patients
on a continuous basis; the remaining 160 patients did not
receive anticoagulant therapy or received it only for several
months. The diagnosis of ischemic disease was based on
clinical examination. Contrast arteriographic studies,
generally of four major cerebral vessels, were performed on
selected patients to identify the site and completeness of the
obstructive process.
163
The follow-up period ranged from three to eight years for
both groups. Information is not available in this particular
report about recurrences of TIA, but a preponderance of the
cerebral infarcts appeared in the control group — 33 as compared with four in the treated group. The total number of
deaths was similar in the two groups. Information is incomplete about bleeding problems; the authors stated that
the incidence of fatal cerebral hemorrhage was higher in the
group treated (7.4% versus 4.4% in the untreated patients)
but they indicated that this increase was minimal compared
with the incidence of fatal thrombotic infarcts.
The major deficiencies of the study are the lack of randomization and the failure to define the criteria for treatment, nontreatment, and curtailment of treatment. For example, it could have been that the anticoagulant treatment
was withdrawn from any patient who became very ill.
Should patients such as these be counted in the control
group, then clearly the evaluation of efficacy would be biased
in favor of anticoagulant therapy.
Whisnant, Matsumoto and Elveback86 carried out a chart
review of patients in the Rochester community during a 15year period. Of 198 patients identified as having experienced
TIA, 80 had received long-term anticoagulant therapy and
118 had not. There was no known selection process for those
who were treated with anticoagulants; the levels of blood
pressure were generally similar in the two groups. The actuarial method of analysis was used to compare the survival
times of patients treated and those not treated with anticoagulants. These, in turn, were compared with survivorship of a group of the same age and sex, using the Minnesota
death rates for 1960. The conclusion was reached that no
difference in age-corrected survival between treated and untreated patients was apparent, but that there were
significantly fewer strokes in treated patients. It is interesting that this difference was established within a month
after starting therapy and that no significant difference
between treated and untreated groups was seen during the
remainder of the five-year follow-up period.
The criticism of the study is, once again, that patients
were not randomized. The authors were interested in the
probability of stroke following the first TIA and did not
document the incidence of further TIA in their paper. One
could raise the question (common to all the nonrandomized
studies) as to whether some of those patients who were
untreated and had prognostic factors which made
anticoagulation inappropriate or who died while under
evaluation for possible anticoagulant therapy would be listed
as untreated subjects, and hence introduce an obvious bias in
favor of the treated group.
Friedman et al.35 reported an epidemiologic survey of a
retirement community in which 60 cases of TIA were
observed over a period of five years. Most of these patients
were neither hospitalized nor referred beyond their primary
physicians and the neurologic examinations therefore were
less uniformly thorough than in most reported studies. Of
these 60 individuals, the subsequent clinical course was examined in 44, half of whom had received anticoagulants. It
was observed that only one subsequent cerebral thrombosis
was seen among the 22 patients receiving anticoagulants,
whereas seven patients in the untreated group had cerebral
thrombosis.
Once more, these observations were not obtained from a
164
STROKE
Downloaded from http://stroke.ahajournals.org/ by guest on June 18, 2017
controlled experiment. Information is not given as to why
some were treated with anticoagulants and some were not.
Also, the treated and untreated groups may have been quite
different as regards important clinical characteristics and
the manner in which they were managed; for example, the
patients receiving anticoagulants were noted to be generally
somewhat younger (68 years) than those for whom therapy
was not elected (72 years).
Some of these studies, including that of Fisher,64 raise a
question about the effect on TIA when anticoagulants are
withdrawn from a patient who has been receiving them for
some time. Marshall and Reynolds67 addressed themselves
to this problem. Of 26 male patients who had been receiving
anticoagulants for at least three months because of having
previously experienced TIA, one-half had therapy gradually
replaced by identical placebos. The allocation of patients to
continued therapy or withdrawal was made according to a
random process after matching patients in pairs for site*erf
the ischemia (carotid or vertebral-basilar) and previous
duration of the anticoagulant therapy. These investigators
found that only two of the 13 patients who continued anticoagulants had further TIA over a follow-up period of four
months, compared with eight of the 13 in the group for
whom anticoagulant therapy was withdrawn. This difference
is statistically significant (p < 0.05).
Summary of Clinical Studies of Anticoagulant Therapy in
TIA
Of the eight studies in which the effect of anticoagulants
on TIA has been evaluated, four were randomized and four
were not. Most of them included only a small number of
patients and for three studies, the average duration of
follow-up was only about 12 months. The consistent finding
in all eight was that no benefit derived from anticoagulants
in terms of lower mortality. However, the report by Whisnant, Matsumoto and Elveback66 did show significantly
fewer strokes in the 80 treated patients as compared with the
160 patients who did not receive anticoagulant therapy.
Significantly, none of the four randomized trials showed a
reduced incidence of stroke attributable to anticoagulant
therapy, whereas all four nonrandomized studies indicated
that there was such a benefit.
Three of the eight studies did not report on recurrences of
TIA. In three other studies, including two which were randomized, the investigators concluded that treatment with
anticoagulants resulted in a reduction of TIA. Of these,
Fisher's study64 was neither randomized nor blind and the
apparent benefit of anticoagulants may simply be a reflection of other factors. The Veterans Administration
Cooperative Study61 was not blind, and the authors admit
that this factor could have influenced their interpretation of
recurrent symptoms. Baker, Broward, Fang et al.45 showed
an appreciable difference between the two treatment groups
in the number of patients having attacks in the first month as
well as a remarkable difference in the total number of attacks (although most of this was chargeable to only two
patients). The study was not blind, thus weakening the inference of actual benefit.
Baker, with Schwartz and Rose,63 subsequently carried
out a study with larger numbers and a longer period of
follow-up. Although they concluded that the data suggested
VOL 8, No
1, JANUARY-FEBRUARY
1977
a reduction in TIA with anticoagulants, they did not focus
attention on recurrences which took place after patients
stopped therapy. These recurrences cannot necessarily be
disregarded. If included in the overall evaluation, the results
for the two treatment groups are remarkably similar.
Finally, although the numbers of patients were small, the
best study methodologically is that of Pearce and his
colleagues62 who were unable to demonstrate any benefit of
anticoagulants in reducing TIA.
The report of Marshall and Reynolds67 is interesting in
that it showed an increased incidence of TIA after withdrawal of anticoagulant therapy. The question arises as to
whether this represents a protective effect of anticoagulants
or simply a rebound phenomenon after withdrawal of
therapy.
In summary, therefore, the evidence is unconvincing that
anticoagulants are of benefit to patients with TIA in terms
of increased survival and reduced incidence of stroke.
Further, the evidence that anticoagulants are of benefit in
reducing the number of TIA recurrences remains equivocal.
Even if there were some reduction in the frequency of TIA,
one must ask if this is really a benefit when the four randomized trials showed that 16 persons in the treated group
died, compared with only ten in the control group, and correspondingly there were 13 major bleeding problems in the
treatment group compared with only four in the control
group.
Randomized Trials of Anticoagulants in Patients With
Progressing Stroke
(Stroke-in-Evolution)
Carter43 processed 122 patients with progressing ischemic
stroke out of 277 consecutive admissions of patients with
cerebral thrombosis. Of these, 76 were ultimately selected as
eligible for study after excluding those with severe hypertension, xanthochromic or blood-stained CSF, evidence of
hepatic disease or past peptic ulceration, age more than 70
years, and other conditions. These 76 patients were allocated
at random to treatment or control. In another account of
this study, Carter48 indicated that, "Alternate patients were
treated by anticoagulants . . . " and so a question arises as to
whether or not patients were truly randomized. The treatment was begun with heparin and followed for four weeks
with oral anticoagulants (phenindione) before treatment was
slowly phased out. The control group did not receive a
placebo and the study, therefore, was not blind. Patients
were followed up for six months. Assessment was made by
clinical examination and the condition of the patient was
classified as recovered, improved, not improved, or died.
The outcomes for the patients were reasonably similar in the
two groups; 11 of the 38 in the control group recovered completely compared with 13 of 38 on anticoagulants. There
were seven deaths in the untreated group compared with
three in the treated group. No information is given as to
when the deaths occurred but autopsies were obtained in five
of the untreated and three of the treated patients.
Pulmonary infarction accounted for three of those in the untreated group. Whether any died as a direct result of their
cerebral infarction is not clear.
When the data were reanalyzed by classifying the patients
as to whether or not the stroke was complete at the time
treatment was begun, those in whom the stroke was still
CEREBRAL ISCHEMIA//oinf Committee for Stroke Resources
Downloaded from http://stroke.ahajournals.org/ by guest on June 18, 2017
progressing did somewhat better when treated with anticoagulants than did the .controls. The appropriateness and
validity of this assessment can be questioned since treatment
would have begun at varying times after randomization in
those given anticoagulants, and no dummy treatment was
given to controls. The nonblindness in the clinical
assessments is also a cause of concern; the author indicates
that, "Selected material is always s u s p e c t . . . "
Baker, Broward, Fang et al.45 included in their series 128
patients with stroke-in-evolution. In the control group, 17 of
67 patients died, compared to 13 of 61 patients in the treated
group. These rates are reasonably similar when allowance is
made for the corresponding periods of follow-up of 15 and
12 months, respectively. Progression of infarction was
recorded in 21 patients in the control and in only eight
patients in the treated group. There were 26 separate
episodes of progression in the controls (14 severe, seven
moderate, five mild), and nine in the treated group (three
severe, three moderate, three mild). Little detail was given
regarding recurrences of TIA except that the number of
monthly periods in which they occurred totaled 20 in the
control group compared with only three in the treated group.
Minor hemorrhagic complications were much more frequent
in the treated group, but the incidence of major complications was low and similar in both groups.
Nonrandomized Trials of Anticoagulants in Patients With
Progressing Stroke
Fisher64 reported 14 cases of progressing stroke. These
patients were treated with heparin followed by oral anticoagulants, apparently for an indefinite period. The results
were evaluated in comparison with a similar group of 14
patients who were not given anticoagulants — the basis on
which these patients were selected from the total group being
unclear. The author was impressed with the improvement
seen when anticoagulants were begun and also with the
deterioration that occurred in three of five persons following
termination of the anticoagulant therapy. He indicated,
however, that conclusions regarding the value of anticoagulants in this group were difficult to draw without control studies.
Millikan and colleagues,41'6Si a as part of their assessment
of anticoagulant therapy, followed 241 patients with
progressing stroke for 12 months. Of the 181 who received
anticoagulants only 12 died, compared with 25 out of 60
patients who were in the control group. This difference is
striking, but the results are difficult to evaluate. That this
finding has potential importance is reflected in the three
separate reports of the study. However, it was neither randomized nor blind. Questions can be raised regarding the
criteria for allocation to treatment and nontreatment, and
how the patients in the two groups compared with respect to
important prognostic factors.
Summary of Clinical Studies of Anticoagulant
Progressing Stroke
Therapy in
The evidence that anticoagulants are of significant benefit
to patients with this condition is quite suggestive but not
conclusive. All four studies were interpreted by their authors
as having demonstrated benefit of treatment, but
165
methodologic weakness in the design raises questions about
the weight to be given the conclusions. The previously stated
limitations of Fisher's64 nonrandomized and nonblind study
weaken the inferences made regarding efficacy. Lack of information about certain selection factors does not permit
firm conclusions to be drawn from the Mayo Clinic study.41
Carter 43 - 48 did not consider stroke as an end point but instead developed a classification of patient function which
would appear to be subjective and therefore open to the
possibility of bias, since the study was nonblind. He also
considered a number of possible reclassifications of his data,
some of which showed a benefit of treatment to be
marginally significant.
The evidence from the study by Baker et al.45 is probably
the most persuasive. There was little difference in the
number of deaths in the two groups but a marked benefit of
anticoagulant therapy with regard to progression of infarction, particularly after the first month of treatment. However, confirmation of these findings would be very desirable.
Randomized Trials of Anticoagulants in Patients With
Completed Stroke
Marshall and Shaw69 carried out a study to assess the influence of anticoagulant therapy on immediate mortality
(within six weeks of the cerebrovascular accident) in patients
under 70 years of age who, during the 72 hours before admission to hospital, had sustained a completed stroke for
which neither a hemorrhagic lesion nor cerebral embolism
was considered to be the causative factor. All patients
qualifying for the trial had lumbar puncture and cerebral
angiography performed, and they were randomized in pairs
to either anticoagulant therapy or control. Treatment was
continued for 21 days and then gradually withdrawn, and
satisfactory anticoagulation was monitored daily using the
Quick one-stage test. The control group did not receive
placebo therapy but otherwise was given the same general
care and physiotherapy as the anticoagulant therapy group.
The lack of placebo therapy and the nonblindness were less
important here, since the patients were hospitalized and
death was the principal end point assessed.
The study comprised 51 patients randomized in pairs to
anticoagulant therapy or control. Twenty-six patients were
treated and 25 served as controls. The two groups were comparable with respect to age, sex, and anatomic site of the
vascular lesion; of ten patients with a mean diastolic
pressure of 110 mm Hg or more, four were in the treated
group and six in the control group. The data on deaths were
analyzed sequentially using a "restricted procedure" to such
a point that it was impossible to conclude in favor of anticoagulant efficacy. In absolute terms, six treated patients
died within six weeks compared with three patients in the
control group. It was decided to continue observation of the
patients up to six months and during this extended period of
follow-up two more patients in the treated group and four in
the control group died. The authors concluded " . . . that
anticoagulants, when used according to the criteria of selection and management of the trial, are not of value."
The Veterans Administration Study61 included 118
patients with a completed stroke who were randomized into
their respective treatment groups. The incidence of TIA and
of stroke were similar in the two groups and there were
166
STROKE
Downloaded from http://stroke.ahajournals.org/ by guest on June 18, 2017
rather more deaths among the treated, 12 compared to
seven, which becomes more significant when account is
taken of the longer period of follow-up for the control
patients. Hemorrhagic complications were far more common in the treatment group than in the control group.
Baker, Broward, Fang et al." included in their study 132
patients who had a completed thrombotic stroke. Fifteen of
the 60 patients in the control group and 18 of the 72 in the
treated group died. When allowance is made for the
difference in exposure in the two groups, the mortality per
1,000 patient-months was 15.4 in the control group and 23.0
in the treated group. Progression of infarction occurred in
five cases in the control group compared to 12 in the treated
group. Little information is given about recurrences of TIA
other than that the number of monthly periods in which such
attacks occurred was four for the control group and two for
the treated group. Once again, there were considerably more
hemorrhagic complications in the treated group than in the
controls, amounting to a total of 31 in the treated as contrasted to two in the control group. Overall, the treated
group fared less well than the control group with respect to
all the important clinical end points.
Hill, Marshall and Shaw70 carried out a trial of long-term
anticoagulant therapy in patients under 70 years of age who
had one or more disturbances of neural function lasting
more than 24 hours and attributable to nonhemorrhagic
arterial disease. Admission to the trial was permitted any
time after 14 days had elapsed since the last acute episode.
Patients were excluded for the usual contraindications to
anticoagulant therapy and were included only if it seemed
likely they would give adequate cooperation.
Cerebral angiography was carried out in about one-third
of the patients entered into the study. Those admitted to the
trial were differentiated by sex and allotted by pairs to a
high-dosage (treatment) or low-dosage (control) group at
random. The former received tablets containing 50 mg
phenindione in a dosage sufficient to maintain the prothrombin time at two to two-and-one-half times the control value.
The low-dosage control group received tablets apparently
identical but containing only 1 mg of phenindione; their
management did not differ otherwise from that of the highdosage group. All patients were seen and blood taken for
prothrombin estimation at least once every four weeks.
After 20 months (end of Phase I), the authors reported their
data and modified the protocol by excluding hypertensive
patients before going on to complete Phase II of the study.
The results of Phase I are summarized in table 5. Not only
did these data provide no evidence of benefit from anticoagulants, but there was some indication that anticoagulants were dangerous, since none of the control group
died, but four deaths occurred in the treated group, all of
which were thought to be attributable to the treatment. One
additional patient died as the result of a possible complication of- treatment. Three of the four patients who died of
cerebral hemorrhage had diastolic blood pressures greater
than 110 mm Hg and it seemed important, therefore, that
the hypertensive patients should be excluded.
Accordingly, the trial was modified by withdrawing the
patients with hypertension. The 94 patients remaining, who
were not hypertensive, continued treatment without interruption according to the original protocol, and an additional 37 patients were admitted over the next year, mak-
VOL 8, No
1, JANUARY-FEBRUARY
1977
ing 131 in total. No more patients were admitted, but
follow-up of those already in the study was continued for
another ten months.
The results for the second phase are also summarized in
table 5. Of the 131 patients in the modified trial, 66 were in
the high-dosage group and 65 in the low-dosage group; these
were comparable in respect to age, sex, and diagnosis of the
anatomical site of the lesion; the average duration of followup was similar for both groups.
The recurrence rate of nonfatal cerebrovascular accidents
differed little in the two groups. The high-dosage group had
17 recurrences in 12 patients, and the low-dosage group 18
recurrences in ten patients. However, the treatment group
had five fatal strokes compared to only one in the controls.
Three of the patients who died while on treatment were
autopsied, and in every case, the cause of death was found to
be cerebral hemorrhage. Another death was attributed to
treatment. Deaths from other causes were six in the treated
group to three in the controls. The overall number of deaths
was 12 in the treatment group and only four in the control
group.
Clearly all the evidence from this study indicates that anticoagulants are of no benefit to patients who have had a completed thrombotic stroke. In fact, it suggests that it is
dangerous to employ anticoagulant treatment for these persons. This trial was extremely well designed and apparently
was carried out with considerable attention to detail. However, by current standards, the level of anticoagulation
achieved might be considered greater than recommended at
the present time. A level of two to two-and-one-half times
the control value was maintained.
Howard et al.,71 as part of a larger investigation, carried
out a controlled double-blind study in which 30 carefully
selected patients with completed thrombotic stroke were
randomized into two groups of equal size, one treated with
anticoagulants and the other with matching placebo, and
followed for 12 months. No information was given regarding
the extent to which anticoagulant control was achieved. No
assessment was made of either TIA or stroke recurrence but
there were three members of each group who died. No major
hemorrhagic complications were encountered and only three
minor complications in each of the two groups. This study
appears to have been well carried out but the numbers of
patients are too few to add significant weight to the accumulated evidence.
McDowell and McDevitt72 carried out a long-term study,
with 92 patients randomized to anticoagulant therapy and
99 to the control group; they were followed for an average of
nearly three years. All participants met well-defined inclusion and exclusion criteria. The control patients received
placebo therapy but were seen less frequently for follow-up
and whether the follow-up clinical assessments were blind is
not clearly indicated. The occurrence of important end
points such as TIA, stroke, death, and hemorrhagic complications was reported within three categories of patients:
those in the control group, those in the treatment group still
on therapy at the time of the episode, and those who had
been on anticoagulants originally but whose medication was
stopped at the time of the attack. The authors chose not to
incorporate the third group in making their evaluation of the
efficacy of anticoagulants and concluded that such therapy
had provided a real benefit.
CEREBRAL ISCHEMIA/Joint Committee for Stroke Resources
Downloaded from http://stroke.ahajournals.org/ by guest on June 18, 2017
However, the matter of discarding one segment of the
study raises an important question concerning the authors'
conclusion. As discussed earlier, once a patient has been randomized to anticoagulant therapy, the subsequent clinical
episodes should be charged against such therapy unless one
can find a specific reason for not doing so. Only in this way is
one given a true picture of what is likely to happen if anticoagulants are recommended as therapy in a defined population. The outcomes relating to all the patients randomized to
treatment and to control are summarized in table 5. One can
observe that the two groups are not appreciably different
with respect to the occurrence of TIA, stroke, and particularly death. As might be expected, bleeding complications in the treated group exceed those in the controls.
Enger and B^yesen73 carried out a well-designed and wellexecuted study in which specific inclusion and exclusion
criteria were clearly defined. Rejected from the study were
those patients who were aged more than 75 years, were unable or unwilling to cooperate, repeatedly exhibited diastolic
blood pressures more than 120 mm Hg, were shown to have
hemorrhagic or xanthochromic cerebrospinal fluid, or were
severely ill from other causes.. All patients were examined by
angiography, and 111 patients with completed thrombotic
stroke were assigned at random to either anticoagulant
therapy or to a matching placebo. The trial was double-blind
and both series were submitted to the same medical and
social care. Follow-up neurologic assessments were made
without the examiner's knowledge of which treatment the
patient had been receiving.
Eleven patients who were followed up for less than three
months were excluded for reasons unrelated to their cerebrovascular disorder. This left 51 remaining in the treated series
and 49 in the controls; initially the groups were comparable
with respect to important clinical characteristics. Anticoagulant control and assessment were satisfactory. After
an average follow-up lasting nearly two years for both
groups, little difference in outcome between the two was evident, as shown in table 5.
After treatment was withdrawn from all patients, those in
both series were followed for an additional 15 months, on
the average. During this time four more cerebral infarctions
appeared in the treated series (one fatal) and five occurred in
the control series, with no fatalities. Thus, no difference was
apparent between the two series with respect to the occurrence of cerebrovascular episodes after discontinuance of
both anticoagulant and placebo administration. The authors
concluded that their study spoke against the long-term use
of peroral anticoagulants in the treatment of patients with a
completed cerebral infarction of a probable thrombotic
and/or atheromatous origin.
Nonrandomized Trials of Anticoagulants in Patients With
Completed Stroke
Howell, Tatlow and Feldman" reviewed the case records
of 272 unselected patients with completed stroke, 77 of
which were discarded either because of death resulting from
the qualifying stroke or because the subject had been
followed for less than one month. Of the 195 patients
remaining, 103 had received anticoagulants and 92 had not.
The average period of follow-up was 16 and 36 months,
respectively, for the two groups. The benefits deriving from
167
administration of anticoagulants appear impressive, in that
28 of the control patients had another stroke compared with
only seven of those who were under treatment; the corresponding numbers for the first 16 months of follow-up
were 13 and five, respectively.
No explanation is given for the marked difference between
the two groups in the length of follow-up period, and, consequently, one must be concerned about the initial comparability of the two groups, especially since the controls
comprised those patients for whom maintenance of anticoagulant therapy would be difficult. The authors indicated
that, "This experience does suggest that anticoagulant
therapy has an effect in controlling the frequency and
severity of cerebral infarction, even if the significance is
limited by our failure to randomize our patients."
Thygesen et al.75 selected into three groups a number of
patients with presumed cerebral infarction. There were 68
for whom anticoagulant therapy was instituted immediately
after admission to the study and who, after six weeks, were
discharged from the hospital as outpatients. Thirty-three of
these continued their anticoagulant therapy; for the remaining 35, therapy had been reduced and was replaced with
placebo after discharge from hospital. The 41 patients in the
third group were given neither anticoagulants nor placebo.
The results for these 41 untreated and the 33 continuously
treated patients were similar. There were four deaths in each
group, 12 patients in each group had TIA, and there were 11
strokes in the treated group compared to seven in the control
group. The 35 patients for whom the anticoagulants were
discontinued and who were discharged from the hospital on
placebo also had similar outcomes.
Considerable care appears to have been taken in selecting
and describing these patients, and all the clinical evaluations
were carried out without knowledge as to which group a particular patient belonged. However, it was not made clear
whether patients were allocated to their respective groups
according to a prescribed randomized arrangement.
Furthermore, it is not evident how the analysis of results
takes account of those patients who were treated with anticoagulants in hospital but could not be followed as outpatients on discharge because they either were too ill or had
died. If these are not counted as failures in the treated group,
then the data reported are likely to present anticoagulants as
having no benefit when, in fact, they might even be harmful.
Summary of Clinical Studies of Anticoagulant Therapy in
Completed Stroke
Of the seven randomized, controlled clinical trials
evaluating anticoagulant therapy in patients with completed
stroke, those reported by Baker, Broward, Fang et al.,45
Hill, Marshall and Shaw,70 and Enger and B^yesen" are
considered to be the most satisfactory in plan and execution.
Each of these contained more than 100 patients who were
randomized into active and placebo therapy groups and
otherwise received similar medical care and attention. The
studies by Hill et al.70 and by Enger and B^yesen73 were both
double-blind and the Baker et al/ 5 study was single-blind.
The findings are consistent and clear that no therapeutic
benefit from anticoagulants derives to patients with completed stroke. Indeed, the study by Hill et al.70 suggests that
actual danger attaches to the treatment of patients such as
168
STROKE
these with anticoagulants. While the design and execution of
the other four randomized trials were relatively less
rigorous, they are reasonably sound and provide further
evidence that no real benefit results from anticoagulant
therapy in patients with completed thrombotic stroke.
Of the two nonrandomized trials, the one reported by
Thygesen et al.75 came to the same conclusion, whereas
Howell and his colleagues74 concluded that some benefit is
evident. However, this last study is methodologically weak
and therefore the data are unconvincing.
Clinical Trials Involving Patients With Cerebral Embolism
Downloaded from http://stroke.ahajournals.org/ by guest on June 18, 2017
Studies of patients with rheumatic heart disease and of
those with acute myocardial infarction allow assessment of
the efficacy of anticoagulants in the prevention of cerebral
embolism, either initial or recurrent, or in altering the immediate course and outcome after the ischemic event has occurred.
Rheumatic Heart Disease. The single properly randomized trial with rheumatic heart disease patients is that of
Baker, Broward, Fang et al.4S which included only 12
patients who were treated with anticoagulants and 16 controls. Therapy was begun with heparin followed by oral anticoagulants if symptoms had occurred during the week prior
to randomization. Most of the patients were seriously ill and
death occurred "early" in approximately 50% of both
groups. Only two of the treated patients were given anticoagulants for a prolonged period. The recurrence rate and
mortality, both early and total, were similar in the two
groups but the sample size was too small for meaningful
conclusions to be drawn. It was not suggested that progression of the neurologic deficit was associated with anticoagulant therapy.
Owren18 compared 17 patients with rheumatic valvular
disease who were treated with oral anticoagulants following
a first episode of embolization with a similar number who
served as untreated controls. During a total of 90 patientyears' follow-up in the treated patients, only one recurrent
embolic episode was observed compared to 22 major
recurrent episodes in 81 patient-years' observations in the
control patients. Details concerning the method of allocation are lacking in this study, as are data pertaining to any
complications that resulted from therapy.
Szekely10 retrospectively reviewed the medical course of
72 patients with rheumatic heart disease and clinically
diagnosed systemic embolization, which involved the cerebral circulation in 48 instances. In 23 of these, anticoagulants were administered for a total of 58 treatmentyears and there were two recurrent embolic episodes, one of
which occurred following temporary discontinuance of the
drug. The incidence of recurrent emboli was 3.4% per
patient-year compared to a 9.6% incidence per patient-year
in 46 patients with embolization who were not treated with
anticoagulants. The author concluded that a benefit from
anticoagulants was shown in those patients treated after a
first embolus.
This author also reported on the outcome in 30 patients
with atrial fibrillation but without embolization who were
followed for 46 patient-years. Comparison was made with a
more or less similar group of 98 patients not given anticoagulants and followed for 499 patient-years. In the treated
VOL. 8, No
1, JANUARY-FEBRUARY
1977
group, two embolic episodes occurred, compared with 34
among the untreated patients. No mention of bleeding complications is made in the report and a major deficiency of the
study is the lack of information as to the basis upon which
cases were selected for treatment and details concerning the
adequacy of follow-up in each group.
McDevitt76 compared the results in 47 patients with
rheumatic heart disease who were followed while on and off
anticoagulants. Twenty-seven embolic events (five involving
cerebral vessels) occurred in 1,315 patient-months on anticoagulant therapy and 132 episodes (33 cerebral) occurred
during 1,437 patient-months off anticoagulants. Considerable detailed information is given concerning the guidelines for the use of anticoagulants followed in the author's
institution, i.e., criteria for administering or discontinuing
anticoagulants, procedure for follow-up assessments and
possible complications of therapy, particularly on a longterm basis. However, the lack of specific information on
these matters as applied to the patients included in the "onoff" study limits one's ability to interpret these data.
Fleming and Bailey16 retrospectively assessed anticoagulant treatment in 217 patients with mitral valvular disease and a history of embolism during 649 patient-years of
treatment. A total of five emboli occurred, an incidence of
0.8% per treatment-year. The authors concluded that this incidence of recurrence was well below that to be expected in
an untreated group and mentioned that several patients in
the study had systemic emboli soon after the drug was withdrawn. The incidence of bleeding was 2.3% per patienttreatment and in only one case was the hemorrhage fatal.
Adams et al.6 conducted a retrospective analysis of a 20year experience with 84 patients having mitral valvular disease and atrial fibrillation and sustaining a stroke from a
clinically diagnosed embolus. One-half of these were given
oral anticoagulant treatment, initiated within two weeks
after the embolic episode; the other half were untreated. The
groups were not concurrent in that most untreated patients
were observed in a ten-year period that began prior to the
one during which the largest number of anticoagulanttreated patients was admitted.
Significant differences in survival favoring the use of anticoagulant therapy were noted. The differences became apparent early and, in fact, were maximal at six months but
continued for ten years of follow-up. Recurrent embolism
was considered the cause of death in 13 of the untreated and
in only four of the treated patients. The lack of concurrency
in the time periods covered makes comparison of these two
groups difficult. The majority of deaths were due to causes
other than emboli and the greatest difference between the
treated and the untreated was observed during the first six
months. This raises the question whether bias entered into
selection of patients for anticoagulant treatment by excluding from receiving the drug those who were the most
seriously ill and therefore had the poorest prognosis. Only
fatal emboli are mentioned in the report and it would be
valuable to know the total incidence of recurrent embolism.
In several uncontrolled studies, including those by Askey
and Cherry77 and Cosgriff," oral anticoagulant therapy was
administered to patients with cardiac disease, usually
associated with atrialfibrillationand systemic embolization.
Most of these authors concluded that the treatment was
practical, reasonably safe, and effective in reducing the in-
CEREBRAL ISCHEMIA/7o;n/ Committee for Stroke Resources
Downloaded from http://stroke.ahajournals.org/ by guest on June 18, 2017
cidence of recurrent embolization.
Wells'9 evaluated the effect of anticoagulant therapy on
the short-term outcome following an acute cerebral embolic
event. In this retrospective study, 63 embolic events occurred in 53 patients not treated with anticoagulants as compared with 34 embolic episodes in a group of 29 patients who
were given anticoagulant drugs. The patients in the two
groups were reasonably comparable.
Anticoagulants were begun within 24 hours of the embolic
episode in most instances and by 48 hours in all; however, an
unspecified number were started with heparin and the rest
were given only oral anticoagulants. The degree of anticoagulation produced was not stated. Death occurred in 25%
of the untreated as compared to 6% of the treated patients,
and severe permanent disability in 35% of the untreated and
40% of the treated patients. One-third of deaths in the untreated patients appear to have happened during the first 24
hours, at a time before any beneficial action of anticoagulants, if such existed, could be expected to take place
in the patients receiving them. This observation suggests the
possibility of bias in case selection, with avoidance of anticoagulants for the most severely ill patients. Evidence was
presented to indicate that anticoagulants, when given to
patients with red cells in the spinal fluid, worsened the
prognosis although the data presented were inadequate to
document this point. The authors concluded that anticoagulants given soon after an embolic event certainly did
not worsen the outcome and that they might have slightly
reduced mortality and had significantly decreased morbidity. It is difficult to draw sound conclusions from this
nonrandomized, retrospective, and nonblind study in which
the case-selection also may well have been biased.
Carter80 also has been interested in the effects of early
anticoagulation. In an uncontrolled study, the data
suggested benefit of anticoagulants begun early, in that twothirds of 49 patients with cerebral embolism recovered or
improved; in contrast, only one-third of a prior group
treated without the use of anticoagulants made a similarly
good recovery. Based on previous experience, the author
identified a history of bleeding disorder, severe hypertension (diastolic pressure of 120 mm Hg and above), age
greater than 70 years and evidence of bleeding into the
cerebrospinal fluid as contraindications to the use of anticoagulants. Presumably because of an unacceptable incidence of bleeding, the dose of anticoagulants was reduced;
however, the revised regimen as outlined still resulted in
levels of anticoagulation which would be considered excessive today. Later, Carter 13 summarized experience gained
with more than 100 patients with cerebral embolism, treated
in various ways. Most of these had either rheumatic heart
disease or ischemic heart disease and atrial fibrillation.
One-third of 34 patients who were untreated died early
from the immediate effects of the embolism (more than onehalf were dead within one year) and more than one-half of
the survivors remained permanently disabled. By contrast,
during the period 1954 to 1957, 43 patients were admitted,
all of whom received a single four-week course of anticoagulants immediately after their embolism. At the end of
six weeks, nine of these had died, eight had made little
recovery, seven were improved, and 19 had recovered. Only
patients with incomplete lesions were admitted to the series
from 1958 to 1963 inclusive. Of 26 patients in this group who
169
were treated with anticoagulants, one died, three were not
improved, five were improved, and 17 had recovered.
The author concluded that anticoagulants are useful in the
immediate treatment of patients with mild or incomplete
lesions. Caution was urged in prescribing treatment for
patients with completed lesions for whom anticoagulants are
of little value and even appear to be hazardous. Late
recurrence of embolism in treated patients was 16% during
two years of observation as compared to 57% in an earlier
untreated group during a similar length of time. Carter
recommended that for patients with embolism, anticoagulant treatment be continued for at least a year, by
which time the period of greatest risk for recurrence will
have passed, particularly for patients with ischemic heart
disease. In those with mitral stenosis, especially when
associated with atrial fibrillation, two years of treatment
were advised and even longer for those in whom embolism
recurred during follow-up.
Myocardial Infarction. Numerous studies have evaluated
the effectiveness of anticoagulant drugs in the management
of patients with acute myocardial infarction and have
resulted in much controversy. Consequently, their
therapeutic role in the routine management of patients such
as these is presently unclear.
Without question, mural thrombus overlying infarcted
myocardium is a frequent source of the material causing
cerebral embolism. Several studies have provided both
clinical and postmortem documentation to suggest that oral
anticoagulants are of value in reducing the incidence of
mural thrombi and consequently of cerebral embolism.
In a study reported by Wright et al.,18 in which more than
1,000 patients were included (442 in the control group and
589 in the treated group), the clinical incidence of cerebral
embolism was approximately 4.9% in the controls and 0.7%
in the treated patients. Similar data were obtained in two
later trials which were randomized, controlled, well
designed, and well executed. These two studies, one sponsored by the Medical Research Council (M.R.C.) 81 and the
other by the Veterans Administration (V.A.),82 reported an
incidence of cerebrovascular lesions diagnosed clinically in
2.5% to 3.7% of patients in the control series, while in the
treated groups the incidence fell to about 1%. In those who
came to autopsy, the incidence was about 7% in controls and
4% in the treated group in the M.R.C. study and 0% in the
treated group of the V.A. study. The presumed source of
embolic material, i.e., mural thrombosis, was found in
nearly one-half of the autopsied patients in the control group
and in 22% of the patients who had received oral anticoagulants.82
Evidence has been presented suggesting benefit of longterm anticoagulants on incidence of cerebral embolism in
patients following recovery from myocardial infarction. For
example, in the study by Harvald et al.,83 a 1:11 difference
between the anticoagulated and control patients was
observed. The cerebral thromboembolic complications were
listed under the heading of "cerebrovascular thrombosis."
Summary of Clinical Trials Evaluating Anticoagulants
Cerebral Embolism
in
Although optimally designed studies on this important
subject are lacking, the data reported here represent more
170
STROKE
Downloaded from http://stroke.ahajournals.org/ by guest on June 18, 2017
than 500 cases. The results have been consistent and suggest
that in patients with rheumatic or ischemic heart disease
anticoagulants reduce the incidence of initial and recurrent
cerebral embolic episodes. If these results are accepted as a
basis, it seems reasonable to recommend that anticoagulant
therapy be employed in such patients who have suffered
cerebral embolism or in those thought to be at high risk of
experiencing such an event. Bleeding is the only significant
complication of prolonged therapy with the oral anticoagulants and the incidence is acceptably low if patients
with systemic hypertension or previous bleeding tendency
are excluded.
The duration of therapy must be individualized and anticoagulants should be discontinued when the risk of further
embolic episodes becomes minimal. In patients with emboli
following myocardial infarction, the recommendation by
Carter 13 that treatment be discontinued after one year seems
reasonable since it is to be expected that organization of the
mural thrombus would have occurred during that time and
further progression of thrombosis would be unlikely. Obviously, if a patient has ventricular aneurysm and repeated
embolism, prolonged therapy, perhaps for an indefinite
period, would be indicated.
With rheumatic heart disease, especially that involving
the mitral valve and with atrial fibrillation, the occurrence of
a systemic embolus probably identifies the palient who will
remain at continuous risk of recurrence and for whom lifelong anticoagulant therapy should be considered.
A more difficult question to answer from the available
data is when anticoagulants should be instituted for a patient
with cerebral embolism. Obviously, to prevent early
recurrences, it is desirable to begin therapy without delay as
soon as it becomes safe to do so. On the other hand, a feeling
of concern is always present that beginning anticoagulant
administration too soon may lead to hemorrhage into the infarcted area.
This concern has its source in studies such as those
reported by Sibley et al.84 and by Moves et al.85 who experimentally produced cerebral infarcts in dogs and
demonstrated that anticoagulation significantly increased
hemorrhage into the infarct. In animals anticoagulated prior
to embolization, the infarct may have been smaller, but was
also more hemorrhagic. When large doses of anticoagulants
were begun on the day of embolization, striking evidence of
intracerebral hemorrhage occurred. When the drugs were
employed in doses of more conventional size, the
hemorrhage that developed was less, but still significantly
more than in the control animals. This tendency to
hemorrhage persisted even when initiation of anticoagulation was delayed for up to three days following cerebral infarction, nor was a reduction in infarct size apparent.
Whisnant85" amplified a thoughtful discussion of the
significance of the data as they relate to the management of
patients and suggested that these observations should not
necessarily militate against the use of anticoagulants in cerebral embolism but that it might be hazardous to start administration within three days of an embolic episode.
Results of clinical studies are inadequate to establish
whether immediate anticoagulation with heparin is hazardous or beneficial to the patient with cerebral embolus. The
studies by Wells79 and Carter 13 suggest possible benefit
VOL 8, No
1, JANUARY-FEBRUARY
1977
without hazard, although selection bias may have influenced
their results, as previously discussed. These authors' conclusion seems reasonable, that an embolic stroke which is completed at the time a patient is seen is unlikely to be benefited
by anticoagulants, and suggests that early heparinization
should be used only for patients with incomplete, progressing symptoms, and then only with caution. The use of
heparin is contraindicated for hypertensive patients and for
those with a tendency toward bleeding. From the standpoint of preventing recurring embolism, it is likely that anticoagulation can be delayed with safety for several days.
Although Daley et al.14 found a high percent of recurrences
occurring in the first week, most studies indicate recurrence
within the first ten days to be uncommon.
Thus, present evidence suggests that patients with cerebral emboli should be given anticoagulant therapy unless a
serious contraindication exists. In those with evidence of
progressing stroke, the therapy should be started early with
heparin. In other patients, treatment can begin with oral
anticoagulants at the time the patient is seen, with
therapeutic anticoagulation being achieved three or four
days later. If, during that interval, evidence of progressing
recurrence develops, heparin should be given.
Platelet Function-Suppressant Therapy
Of the large number of drugs that alter platelet function
tests in vitro, only a few are potentially useful clinically
based upon consideration of toxicity and the dosage required. These drugs, which are available and in therapeutic
use for other of their actions, include acetylsalicylic acid,
clofibrate, dipyridamole, and sulfinpyrazone.
Four prospective studies have been carried out to evaluate
therapy in cerebrovascular disease with drugs that have been
demonstrated to be platelet suppressive. One retrospective
study and some well-documented individual case studies
have also been reported.
Acheson, Danta and Hutchinson86 evaluated the effect of
dipyridamole at two dose levels in a total of 169 patients,
each of whom had partially or completely recovered from a
clinical episode of cerebral ischemia. A dose of 400 mg daily
was used for periods varying from six to 24 months, and this
dosage was later doubled when it became clear that the
lower dosage was of no benefit. Dosage at the higher level
likewise gave no evidence of reducing the frequency of TIA,
ischemic stroke, or death. However, it is unlikely that a
definite answer could have been obtained from this trial in
any event, since the number of patients investigated was
small, the length of study relatively short, and the incidence
of the two most important end points, stroke and death, was
low (table 6).
Acheson and Hutchinson87 compared the effects of
clofibrate with placebo in patients with hypercholesterolemia and cerebral ischemia manifested by either
TIA or ischemic stroke. The study was designed to determine the effect of reducing elevated blood cholesterol on the
incidence of further episodes of cerebral ischemia and the
mortality rate. Although the study design appears adequate,
specific details of randomization and assurance of blindness
with respect to treatment are not provided. The study was
continued for seven years, with an average follow-up of four
171
CEREBRAL ISCHEMIA//omf Committee for Stroke Resources
TABLE 6 Randomized Clinical Trials of Plalelel-Suppressanl Drugs
Study
Acheson et al.86
Acheson, Hutchinson87
Evans
88
Blakely, Gent"
Year
of
report
Drug
(daily dose)
1969 Dipyridamole
(400 mg)
Dipyridamole
(800 mg)
1972 Clofibrate
(1-2 gm)
1973 Sulfinpyrazone
(800 mg)
1975 Sulfinpyrazone
(600 mg)
Condition
No. pts.
T
C
Average
follow-up
period
(mo.)
TIA l(%)
T
c
Results
CVA"
T
c
Death
T
(%)
c
TIA/CVA
77
76
14
38
43
6
9
10
8
TIA/CVA
62
66
11
16
26
11
6
8
5
TIA/CVA
47
48
60
17
15
49
46
49
42
Amaurosis
fugax
CVA
20
20
2.7f
—
—
—
—
43
56
—
—
—
30J
56J
0.55f
6-wk
crossover
—
Up to
' 4 yr
•Strokes.
tMean number of eye symptoms per week.
{Based on estimated four-year death rates from vascular causes.
Downloaded from http://stroke.ahajournals.org/ by guest on June 18, 2017
years per patient. Although the blood cholesterol was
reduced, no differences were noted in the three end points examined, i.e., TIA, stroke, and death (table 6).
Evans88 carried out a double-blind crossover study in
which each of 20 patients with amaurosis fugax, including
six who had TIA also, received either sulfinpyrazone (200
mg four times daily) or an identical placebo, and after six
weeks were then crossed over to the alternative therapy for
another six weeks. Five patients showed improvement on
both treatments and two did not improve on either, but the
remaining 13 patients all showed a significant improvement
on sulfinpyrazone and none on placebo. The difference is
statistically highly significant (p < 0.001). The study is
limited in that it does not include the more important longterm end points of stroke and death but deals only with the
short-term relief of symptoms, although the author indicated that in a few patients followed for periods of six to
18 months, sulfinpyrazone continued to suppress symptoms.
In addition, it is noteworthy that in a current long-term multicenter trial of TIA,89 no single center has produced 20
cases of amaurosis fugax in four years of study. This, therefore, raises some questions about the diagnostic screening of
patients in Evans' study (table 6).
Blakely and Gent90 reported on a double-blind study of
291 institutionalized elderly males of average age 72 years
who were randomly allocated to sulfinpyrazone (600 mg per
day in three divided doses) or placebo, and followed for up
to four years. In this study, 99 patients had a history of prior
stroke, with or without coronary or peripheral arterial disease. Death was the only end point recorded, and the cause
of death was assessed by observers unaware of the treatment
the patient had received. Cumulative survival in the two
treatment groups was compared, using a lifetable analysis.
Total mortality was significantly reduced, in particular,
death from vascular causes in atherosclerotic subjects. In
patients treated with sulfinpyrazone, the estimated four-year
mortality rate was 30%, as contrasted to 56% in the placebo
group. The major criticism of this study is the lack of detail
given about the assessment of cause of death; it would have
been useful to have data on the nonfatal events. One should
note also that these subjects were a specially selected highrisk group and the results may not be applicable to the
general population (table 6).
Dyken et al.91 reported a retrospective study in which were
reviewed all their patients with a provisional diagnosis of
TIA who, over a three-year period, had been admitted to a
clinical vascular ward. Twenty-six of 117 consecutive
patients were identified who satisfied very specific criteria.
The authors noted in retrospect that 15 of these patients
were treated with aspirin (300 mg twice daily), and 11 were
not. There was no difference between these two groups in the
occurrence of ultimate infarction or death. However, only
two (13%) of those treated with aspirin had an additional
TIA over an average follow-up period of 14 months, compared to nine (82%) of those who had no aspirin over an
average follow-up period of 29 months — a highly significant difference (p < 0.001). This study was criticized by its
authors because it was retrospective, and because other unrecognized variables may have accounted for the difference
in number of attacks.
Harrison et al.92 have reported their experience with
aspirin and dipyridamole in two cases of amaurosis fugax.
During a brief follow-up, they found a significantly reduced
number of attacks of transient blindness when the two
patients were treated with aspirin. Dipyridamole was concluded to be ineffective — although the evidence for this
statement is extremely weak. Mundall et al.93 reported a
single case of amaurosis fugax associated with thrombocytosis and spontaneous platelet aggregation in which the
patient's episodes of transient blindness disappeared when
aspirin was given and recurred only when the aspirin was
stopped temporarily. She remained symptom-free over a
follow-up period of five months while on aspirin.
Several studies have been carried out to evaluate the effect
of platelet-suppressive therapy on the incidence of systemic
embolization in patients with prosthetic heart valves. While
such studies do not deal specifically with cerebral embolization, they are important to include in a discussion of trials of
platelet suppressants. The best-designed trial was that
reported by Sullivan et al.94 who compared the effect of
dipyridamole plus coumadin to coumadin therapy alone in a
prospective randomized double-blind trial in 163 patients
with older-type prosthetic heart valves. During a one-year
period of observation they reported a 14% incidence of embolization in those patients receiving coumadin alone, compared to 1.3% in the dipyridamole-coumadin group..These
172
STROKE
data are suggestive that platelet-suppressant therapy may
reduce the emboli from prosthetic heart valves, an increasingly important source.
Studies in Progress
Downloaded from http://stroke.ahajournals.org/ by guest on June 18, 2017
In progress at the present time are two large cooperative
studies evaluating the effects of platelet function-suppressing
drugs in patients with cerebrovascular disease. A Canadian
study," which has been going on for more than four years, is
a multicenter trial comparing the effects of aspirin, sulfinpyrazone, aspirin plus sulfinpyrazone, and placebo in
patients with TIA. The study is designed to determine the
effects of these treatments on the incidence of recurrent transient ischemia and stroke and on survival. A complementary
study also has been started by the same group of investigators in which sulfinpyrazone is being compared with
placebo in patients with TIA who have had reconstructive
cerebrovascular surgery.
A study in the United States, sponsored by the National
Institutes of Health, is comparing the effects of aspirin with
those of placebo in both surgical and nonsurgical patients
with TIA. This study, which has been in progress for approximately two years, is assessing the effect of aspirin on
recurrent TIA, stroke, and survival.
Summary of Trials Evaluating Platelet-Suppressant
Therapy in Cerebrovascular Disease
Gent 95 has reviewed the reported clinical evaluations of
platelet inhibitor therapy in cerebrovascular disease. The
conclusion expressed there remains unchanged after a reevaluation: none of the studies reported so far has established efficacy with regard to the important end points of
stroke and death, although some evidence suggests that both
aspirin and sulfinpyrazone may reduce recurrence of TIA.
The patients in the two Acheson86- " studies were not well
defined clinically and appeared to be a heterogeneous group.
Both studies were limited in terms of total patient-years of
study and neither was likely to produce an answer relating to
the prevention of stroke or to increased survival. Evans'88
study was well controlled but the outcome related only to
short-term relief of symptoms and the nature of these symptoms was uncertain. Blakely and Gent90 studied elderly people who are at high risk of death from both myocardial infarction and stroke and thus are not representative of the
general population. This is, however, an interesting clinical
model and has provided valuable data confirming that initiation of a long-term multicenter trial with sulfinpyrazone
would be a worthwhile endeavor.
The study by Dyken et al.91 was thoughtfully carried out
and analyzed but, as the authors recognized, in a retrospective
study of this kind it is possible that unknown factors other
than aspirin might account for the differences in frequency
of attacks. They stressed the importance of their findings in
strongly supporting the urgent need for prospective controlled studies before platelet-suppressive drugs become
widely adopted and indiscriminately used.
The members of the study group hope that the data from
the ongoing Canadian and U.S. multicenter studies will
provide definitive information as to the role of these platelet
function-suppressive drugs in the treatment of ischemic
cerebrovascular disease.
VOL 8, No
1, JANUARY-FEBRUARY
1977
Thrombolytic Therapy
Because brain tissue is so extremely sensitive to hypoxia,
the disastrous consequences of thrombotic occlusion of cerebral vessels result almost entirely from reduction in blood
flow. Irreversible damage occurs within ten minutes of total
cessation of blood flow and within 60 minutes if blood
supply is decreased below 20% of normal. Therefore, the
concept of employing therapy in thromboembolic cerebrovascular ischemia to dissolve the obstructing thrombotic
material, and thereby to restore blood flow rapidly while
salvage is still possible, is theoretically a very attractive one.
Such treatment would permit removal of single or multiple
thrombi from extracranial vessels as well as from surgically
inaccessible intracranial vessels. Nevertheless, some potential hazards are associated with this approach. These include
the danger of converting a so-called white infarct to a red or
hemorrhagic infarct when blood flow to ischemic or infarcted tissue is reestablished. In addition, emboli may be
produced from the dissolving thrombi; these could enter
previously uninvolved vessels and cause worsening of
ischemia.
Several approaches are available to produce and sustain a
marked increase in the fibrinolytic activity of blood and
thereby create a "thrombolytic" state. These include the intravenous administration of plasmin which has been activated from the precursor plasminogen by one of several activators, e.g., streptokinase or urokinase. Alternatively, the
plasminogen activators may be infused and produce a
thrombolytic state by activating in vivo plasminogen in the
plasma or, perhaps even more desirably, the plasminogen
which is incorporated on the surface and within the latticework of a thrombus or embolus.96
Experience with thrombolytic agents in the management
of patients with cerebrovascular ischemia is limited. Early
studies with small numbers of patients involved the use of
plasmin-plasminogen activator mixtures and the results
suggest that lysis of angiographically documented cerebral
thrombi is achievable.97-98 Later the use of streptokinase was
shown to be effective in individual cases.99 Meyer et al.100
conducted a nonrandomized and uncontrolled trial including
70 patients with clinically diagnosed stroke-in-evolution.
Treatment consisted of intravenous infusion of urokinaseactivated or streptokinase-activated plasminogen in combination with full-dose heparin for four to six hours daily on
two to four consecutive days. Improvement was observed in
54% of these patients and stabilization without further
deterioration of their condition occurred in another 20%. In
half of the ten cases in which angiography was performed
before and following treatment, resolution of the occlusion
was demonstrated and in two others improvement was
observed. The outcome was impressively better if treatment
was begun within 72 hours after onset of symptoms.
These results were considered sufficiently encouraging to
justify further study and led the same authors to perform a
randomized controlled trial evaluating the effects of
streptokinase-activated human plasminogen.101 Forty normotensive patients with stroke-in-evolution were studied. In
85%, the status of the cerebral circulation was confirmed by
angiography — in most cases carotid or middle cerebral
artery occlusion from thrombus or embolus was
demonstrated. Patients were randomized and treated in a
double-blind manner with either streptokinase-plasmin mix-
CEREBRAL ISCHEMIA//o;>» Committee for Stroke Resources
Downloaded from http://stroke.ahajournals.org/ by guest on June 18, 2017
ture or placebo for four-hour periods on three successive
days, in combination with full-dose subcutaneous heparin,
followed by oral anticoagulants. Treatment was begun
within 72 hours after onset of symptoms of cerebral
ischemia. The effect of treatment on components of the
fibrinolytic system was documented by serial determinations
of fibrinogen and plasminogen levels. Clinical assessment of
results was made by neurologic examinations, performed by
the same neurologist, prior to and ten days following initiation of treatment. In addition, 20% of patients underwent
posttreatment angiography.
The patients appeared to be well matched generally,
although those given lytic therapy were somewhat older than
those in the placebo group. The trial was apparently well
conducted. No significant differences in outcome were noted
between the two groups. Fourteen patients died, seven in
each group, and nine patients in each group improved. Eight
patients underwent follow-up angiography and in one from
each group there was evidence of improved blood flow. Biochemical evidence that lytic activity had been produced was
observed in 80% of those given thrombolytic therapy,
although the intensity of the changes was not detailed in the
report. Complications were minimal. The authors concluded
that no significant differences between the groups had been
demonstrated but considered it likely that the period of
treatment had been too short and that posttreatment angiography of all patients would be required to determine
precisely the effects of fibrinolytic therapy.
Another randomized and double-blind trial102 was conducted which had design features similar to the earlier one
except that all but one of the patients underwent angiography prior to randomization and arteriography was
repeated after ten days in those who survived. Streptokinase was employed as the thrombolytic agent and the drug
was administered by sustained infusion for a single six-hour
period beginning within 72 hours after onset of ischemic
symptoms. Heparin followed by coumadin was given to both
control and treated cases after the streptokinase infusion.
The dose of streptokinase was individualized and determined in a conventional manner. Clinical assessment of
results was made by a neurologist presumably blind to treatment, based upon evaluation prior to and ten days following
admission to the trial.
Seventy-three patients were studied. In the 37 given
thrombolytic therapy, there were 13 deaths (35%) and improvement in 43% compared to an 11% mortality and improvement in 58% of the control patients. Even when the
dose of streptokinase was progressively lowered in an
attempt to avoid hemorrhagic complications, the deleterious
effect continued. Cerebral hemorrhage was demonstrated in
three of the ten patients treated with streptokinase, and
another had a hemorrhagic infarct at autopsy. Furthermore, there were six examples of anemic infarction. In contrast, no cerebral hemorrhage was evident in the three control patients autopsied and each had white infarction.
Evidence of dissolution of thromboembolic material was
observed in all patients who received streptokinase. The data
relating to follow-up angiography are not stated clearly
enough to permit determination as to the frequency of
thrombolysis in the two groups, but this occurred frequently,
though not invariably.
The authors concluded that treatment of stroke-in-
173
evolution with thrombolytic agents was of no apparent value
and, in fact, in the case of streptokinase was harmful to the
patient. They believed that dissolution of thrombus and the
reperfusion of ischemic or infarcted tissue probably explained the frequent cerebral hemorrhage or hemorrhagic
infarction observed.
Thus, from the admittedly limited amount of reported experience, it must be concluded that, although theoretically
attractive, the employment of thrombolytic therapy in acute
cerebral ischemia appears to lack benefit and, as has been indicated by several authors,101"105 may, in fact, be hazardous.
Conclusions
1. Stroke ranks third in frequency as a cause of death in
the USA, accounting for approximately 200,000 deaths
yearly. Of the numerous types of stroke, the thrombosisrelated ischemic variety leads the list by a wide margin. The
economic impact of stroke is estimated to have totaled 6.2
billion dollars in 1972, and is undoubtedly much higher today.
2. As the first essential step in diagnosis, cerebral
hemorrhage should be identified, if it is the cause of the
stroke. The physician should then attempt to distinguish
between cerebral embolism and cerebral thrombosis,
although exact clinical differentiation of the several varieties
of ischemic stroke is not always possible. He should also
attempt to categorize the stroke according to the clinical
condition encountered, i.e., transient ischemic attack (TIA),
progressing stroke (stroke-in-evolution), and completed
stroke.
3. The physician should remain familiar with basic information relative to the mechanisms of blood clotting and with
factors operative when consideration is given to the use of
the various types of antithrombotic therapy in the patient
with, or at risk of, stroke.
4. The study group has outlined in detail a number of
trials conducted in the past, with a critique of each.
5. The information presently available indicates that anticoagulant therapy should be considered for the patient with
ischemic cerebrovascular disease on an individually selective
basis, depending on the pathogenesis of the cerebral lesion,
the type of disorder encountered, and concurrent systemic
disorders. The physician must familiarize himself with the
contraindications to the use of anticoagulant drugs, in particular. If he elects to use anticoagulants, a calculated risk is
introduced for his patient, namely, the possibility of
hemorrhage. Many findings from clinical trials are controversial and can be reconciled only by further definitive
data.
6. Indications for the use of anticoagulants in TIA remain controversial. In reported clinical trials, anticoagulant therapy has been demonstrated neither to increase
survival rate nor to reduce the incidence of subsequent
stroke; the evidence that anticoagulants are of benefit in
reducing the incidence of TIA recurrences is equivocal.
7. In ischemic progressing stroke (stroke-in-evolution),
anticoagulant therapy may have value in retarding the
progress of the infarction.
8. Present evidence appears clear that, in patients with
completed stroke, anticoagulant drugs have not proved to be
useful.
174
STROKE
Downloaded from http://stroke.ahajournals.org/ by guest on June 18, 2017
9. According to current information, embolism of cardiac
(rheumatic or ischemic heart disease) origin can be
prevented with anticoagulants. These drugs should be given
unless the patient has a condition that would contraindicate, such as hypertension or bleeding tendencies. In
most cases of rheumatic heart disease, long-term treatment
is necessary.
10. A review of platelet function-suppressant therapy,
which is a popular subject today, indicates that no clinical
trial so far has established the efficacy of these preparations
with regard to the end points of stroke and death. Future
developments in this field may or may not temper the present enthusiasm which surrounds the use of antithrombotic
agents. To be awaited with great interest are the reports of
the United States and Canadian cooperative studies on the
efficacy of platelet-suppressant therapy.
11. The employment of thrombolytic therapy, although
theoretically attractive, appears to lack benefit and is not
safe.
Acknowledgment
The members of the study group wish to thank the following colleagues for
reviewing the manuscript and providing valuable suggestions and advice.
Alvan R. Feinstein, M.D.
Professor of Medicine and Epidemiology
Yale University School of Medicine
New Haven. Connecticut 06510
William K. Hass, M.D.
New York University School of Medicine
550 First Avenue
New York, New York 10016
Jack P. Whisnant, M.D.
Professor and Chairman
Department of Neurology
Mayo Clinic and Mayo Foundation
Rochester, Minnesota 55901
References
1. Joint Committee for Stroke Facilities: I. Epidemiology for stroke
facilities planning. Stroke 3: 360-371 (May-Jun) 1972
2. Kurtzke JF: Epidemiology of cerebrovascular disease. In Siekert RG
(ed): Cerebrovascular Survey Report for Joint Council Subcommittee
on Cerebrovascular Disease, National Institute of Neurological and
Communicative Disorders and Stroke and National Heart and Lung
Institute (revised). Rochester, Minnesota, Whiting Press Inc, pp
213-242, 1976
3. Cooper BS, Rice DP: The economic cost of illness revisited. Social
Security Bulletin (Feb) 1976
4. Kurtzke JF: Epidemiology of Cerebrovascular Disease. New York,
Springer-Verlag, 1969
5. Wolf PA: Hypertension as a risk factor for stroke. In Whisnant JP,
Sandok BA (eds): Cerebral Vascular Diseases. Ninth Conference. New
York, Grune and Stratton Inc, pp 105-112, 1974
6. Adams GF, Merrett JD, Hutchinson WM, et al: Cerebral embolism and
mitral stenosis: Survival with and without anticoagulants. J Neurol
Neurosurg Psychiat 37: 378-383 (Apr) 1974
7. Hernandez LM: Study of cerebral and peripheral emboli, registered in
the National Institute of Cardiology of Mexico, from September, 1944,
to May, 1946. Am Heart J 33: 712 (May) 1947
8. Harris AW, Levine SA: Cerebral embolism in mitral stenosis. Ann
Intern Med IS: 637-643 (Oct) 1941
9. McDevitt E: Anticoagulant therapy in the treatment of cerebral thrombosis. In Wright I, Millikan C (eds): Cerebral Vascular Diseases. Second Conference. New York, Grune and Stratton Inc, pp 125-145, 1957
10. Szekely P: Systemic embolism and anticoagulant prophylaxis in
rheumatic heart disease. Br Med J 1: 1209-1212 (May 9) 1964
11. Keen G, Leveaux VM: Prognosis of cerebral embolism in rheumatic
heart disease. Br Med J 2: 91-92 (Jul 12) 1958
12. Coulshed N, Epstein EJ, McKendrick CS, et al: Systemic embolism in
mitral valve disease. Br Heart J 32: 26-34 (Jan) 1970
13. Carter AB: Prognosis of cerebral embolism. Lancet 2: 514-519 (Sep 11)
1965
14. Daley R, Mattingly TW, Holt CL, et al: Systemic arterial embolism in
rheumatic heart disease. Am Heart J 42: 566-581 (Oct) 1951
VOL 8, No
1, JANUARY-FEBRUARY
1977
15. Fleming HA, Bailey SM: Mitral valve disease, systemic embolism and
anticoagulants. Postgrad Med J 47: 599-604 (Sep) 1971
16. Owren PA: The results of anticoagulant therapy in Norway. Arch
Intern Med 111: 240-247 (Feb) 1963
17. Browne TR III, Poskanzer DC: Treatment of strokes. N Engl J Med
281: 594-602 (Sep II), 650-657 (Sep 18) 1969
18. Wright IS, Marple CD, Beck DF: Myocardial Infarction: Its Clinical
Manifestations and Treatment with Anticoagulants. A Study of 1,031
Cases. New York, Grune and Stratton Inc, 1954
19. Millikan CH, Bauer RB, Goldschmidt J, et al: A classification and outline of cerebrovascular diseases II. Stroke 6: 564-616 (Sep-Oct) 1975
20. Fisher CM: Observations of the fundus oculi in transient monocular
blindness. Neurology 9: 333-347 (May) 1959
21. Russell RWR: Observations on the retinal blood vessels in monocular
blindness. Lancet 2: 1422-1428, 1961
22. Hollenhorst RW: Significance of bright plaques in the retinal arterioles.
JAMA 178: 23-29 (Oct 7) 1961
23. Barnett HJM: Transient cerebral ischemia: Pathogenesis, prognosis and
management. Ann R Coll Physicians Surg Can 7: 153-173, 1974
24. Hass WK, Fields WS, North RR, et al: Joint study of extracranial
arterial occlusion. II. Arteriography, techniques, sites and complications. JAMA 203: 961-968 (Mar 11) 1968
25. Hass WK: Occlusive cerebrovascular disease. Med Clin North Am 56:
1281-1297 (Nov) 1972
26. Sanders WE: Undulant fever meningitis (organism in spinal fluid). J
Iowa Med Soc 21: 510-511 (Sep) 1931
27. Barnett HJM, Hyland HH: Non-infective intracranial venous thrombosis. Brain 76: 36-49 (Mar) 1953
28. Canadian Collaboration Platelet-Inhibiting Drug Trial. Unpublished
data
29. Reed RL, Siekert RG, Merideth J: Rarity of transient focal cerebral
ischemia in cardiac dysrhythmia. JAMA 223: 893-895 (Feb 19) 1973
30. Brice JG, Dowsett DJ, Lowe RD: Haemodynamic effects of carotid
artery stenosis. Br Med J 2: 1363-1366 (Nov 28) 1964
31. Romanul FCA, Abramowicz A: Changes in brain and pial vessels in
arterial border zones. A study of 13 cases. Arch Neurol 11:40-65 (Jul)
1964
32. Whisnant JP, Matsumoto N, Elveback LR: Transient cerebral ischemic
attacks in a community. Mayo Clin Proc 48: 194-198 (Mar) 1973
33. Drake WE, Drake MAL: Clinical and angiographic correlates of
cerebrovascular insufficiency. Am J Med 45: 253-270 (Aug) 1968
34. Millikan CH: Treatment of occlusive cerebrovascular disease. In
Siekert RG (ed): Cerebrovascular Survey Report for Joint Council Subcommittee on Cerebrovascular Disease, National Institute of Neurological and Communicative Disorders and Stroke and National Heart
and Lung Institute (revised). Rochester, Minnesota, Whiting Press Inc,
pp 141-171, 1976
35. Friedman GD, Wilson WS, Mosier JM, et al: Transient ischemic attacks in a community. JAMA 210: 1428-1434 (Nov 24) 1969
36. Meyer JS, Guiraud B, Bauer R: Clinical and pathophysiological considerations of atherosclerotic and thrombotic disease of the carotid
arteries. In Vinken PJ, Bruyn GW (eds): Vascular Diseases of the Nervous System. Handbook of Neurology. New York, American Elsevier
Pub Co Inc, vol II, pp 327-365, 1972
37. Marshall J, Wilkinson IMS: The prognosis of carotid transient
ischaemic attacks in patients with normal angiograms. Brain 94:
395-402, 1971
38. Marshall J, Meadows S: The natural history of amaurosis fugax. Brain
91: 419-434 (Sep) 1968
39. Marshall J: The natural history of transient ischaemic cerebrovascular
attacks. Q J Med 33: 309-324 (Jul) 1964
40. Ziegler DK: Hypertensive vascular diseases of the brain. In Vinken PJ,
Bruyn GW (eds): Vascular Diseases of the Nervous System. Handbook
of Neurology. New York, American Elsevier Pub Co Inc, vol 11, pp
552-577, 1972
41. Millikan CH: Reassessment of anticoagulant therapy in various types of
occlusive cerebrovascular disease. Stroke 2: 201-208 (Mar-Apr) 1971
42. Jones HR, Millikan CH: Temporal profile (clinical course) of acute
carotid system cerebral infarction. Stroke 7: 64-71 (Jan-Feb) 1976
43. Carter AB: Anticoagulant treatment in progressing stroke. Br Med J 2:
70-73 (Jul 8) 1961
44. Millikan CH: The indications for various methods of treatment of
occlusive cerebrovascular disease. Trans Am Neurol Assoc98:206-218,
1973
45. Baker RN, Broward JA, Fang HC, et al: Anticoagulant therapy in cerebral infarction. Report on cooperative study. Neurology 12: 823-835
(Dec) 1962
46. Hass WK, Goldensohn ES: Clinical and electroencephalographic considerations in the diagnosis of carotid artery occlusion. Neurology 9:
575-589 (Sep) 1959
47. Fisher CM: Anticoagulant therapy in cerebral thrombosis and cerebral
embolism: A national cooperative study, interim report. Neurology 11:
119-149 (Apr) 1961
48. Carter AB: Use of anticoagulants in patients with progressive cerebral
infarction. Neurology 11: 601-609 (Jul) 1961
CEREBRAL ISCHEMIA/Zo/nr Committee for Stroke Resources
Downloaded from http://stroke.ahajournals.org/ by guest on June 18, 2017
49. Harker LA, Slichter SJ: Platelets and fibrinogen consumption in man.
N Engl J Med 287: 999-1005 (Nov 16) 1972
50. von Rokitansky C: Handbuch der Palhologischen Anatomic I Band.
Handbuch der Allgemeinen Pathologischen Anatomic. Vienna,
Braumuller und Seidel, p 510, 1842. Day GE (Eng trans) London,
Sydenham Society, vol 4, p 261, 1852
51. Duguid JB: Thrombosis as a factor in the pathogenesis of coronary
atherosclerosis. J Pathol Bacteriol 58: 207-212 (Apr) 1946
52. Poole JCF, Sanders AG, Florey HW: The regeneration of aortic endothelium. J Pathol Bacteriol 75: 133-143 (Jan) 1958
53. Chandler AB, Hand RA: Phagocytized platelets: A source of lipids in
human thrombi and atherosclerotic plaques. Science 134: 946-947 (Sep
29) 1961
54. Wu KK, Hoak JC: A new method for the quantitative detection of
platelet aggregates in patients with arterial insufficiency. Lancet 2:
924-926 (Oct 19) 1974
55. Kalendovsky Z, Austin J, Steele P: Increased platelet aggregability in
young patients with stroke. Diagnosis and therapy. Arch Neurol 32:
13-20 (Jan) 1975
56. Wu KK, Hoak JC: Increased platelet aggregates in patients with transient ischemic attacks. Stroke 6: 521-524 (Sep-Oct) 1975
57. Yin ET, Wessler S, Stoll PJ: Biological properties of the naturally occurring plasma inhibitor to activated factor X. J Biol Chem 246:
3703-3711 (Jun 10) 1971
58. Smith JB, Willis AL: Aspirin selectively inhibits prostaglandin production in human platelets. Nature (New Biol) 231: 235-237 (Jun 23) 1971
59. Genton E, Gent M, Hirsh H, et al: Platelet-inhibiting drugs in the
prevention of clinical thrombotic disease. N Engl J Med 293:1174-1178
(Dec 4), 1236-1249 (Dec 11), 1296-1300 (Dec 18) 1975
60. Marquardsen J: The natural history of acute cerebrovascular disease: A
retrospective study of 769 patients. Acta Neurol Scand 45 (Suppl 38):
9-188, 1969
61. Baker RN: An evaluation of anticoagulant therapy in the treatment of
cerebrovascular disease: Report of the Veterans Administration
cooperative study of atherosclerosis. Neurology 11: 132-138 (Apr) 1961
62. Pearce JMS, Gubbay SS, Walton JN: Long-term anticoagulant therapy
in transient cerebral ischaemic attacks. Lancet 1: 6-9 (Jan 2) 1965
63. Baker RN, Schwartz WS, Rose AS: Transient ischemic strokes. A
report of a study of anticoagulant therapy. Neurology 16:841-847 (Sep)
1966
64. Fisher CM: The use of anticoagulants in cerebral thrombosis.
Neurology 8: 311-332 (May) 1958
65. Siekert RG, Whisnant JP, Millikan CH: Surgical and anticoagulant
therapy of occlusive cerebrovascular disease. Ann Intern Med 58:
637-641 (Apr) 1963
66. Whisnant JP, Matsumoto N, Elveback LR: The effect of anticoagulant
therapy on the prognosis of patients with transient cerebral ischemic attacks in a community. Rochester, Minnesota, 1955 through 1969. Mayo
Clin Proc 48: 844-848 (Dec) 1973
67. Marshall J, Reynolds EH: Withdrawal of anticoagulants from patients
with transient ischaemic cerebrovascular attacks. Lancet 1: 5-6 (Jan 2)
1965
68. Millikan CH: Therapeutic agents — current status: Anticoagulant
therapy in cerebrovascular disease. In Siekert RG, Whisnant JP (eds):
Cerebral Vascular Diseases. Fourth Conference. New York, Grune and
Stratton Inc, pp 181-184, 1965
69. Marshall J, Shaw DA: Anticoagulant therapy in acute cerebrovascular
accidents. A controlled trial. Lancet 1: 995-998 (May 7) 1960
70. Hill AB, Marshall J, Shaw DA: Cerebrovascular disease: Trial of longterm anticoagulant therapy. Br Med J 2: 1003-1006 (Oct 20) 1962
71. Howard FA, Cohen P, Hickler RB, et al: Survival following stroke.
JAMA 183: 921-925 (Mar 16) 1963
72. McDowell F, McDevitt E: Treatment of the completed stroke with longterm anticoagulant: Six and one-half years' experience. In Siekert RG,
Whisnant JP (eds): Cerebral Vascular Diseases. Fourth Conference.
New York, Grune and Stratton Inc, pp 185-199, 1965
73. Enger E, Bjjyesen S: Long-term anticoagulant therapy in patients with
cerebral infarction. A controlled clinical study. Acta Med Scand 178
(Suppl 438): 1-61, 1965
74. Howell DA, Tatlow WFT, Feldman S: Observations on anticoagulant
therapy in thromboembolic disease of the brain. Can Med Assoc J 90:
611-614 (Mar 7) 1964
75. Thygesen P, Christensen E, Dyrbye M, et al: Cerebral apoplexy. A
clinical, radiological, electroencephalographic and pathological study
with special reference to the prognosis of cerebral infarction and the
result of long-term anticoagulation therapy. Dan Med Bull 11:233-257
(Dec) 1964
76. McDevitt E: Treatment of cerebral embolism. Mod Treat 2:52-63 (Jan)
1965
77. Askey JJ, Cherry CB: Thromboembolism associated with auricular
fibrillation. JAMA 144: 97-100 (Sep 9) 1950
78. CosgrifTSW: Prophylaxis of recurrent embolism of intracardiac origin.
JAMA 143: 870-872 (Jul 8) 1950
175
79. Wells CE: Cerebral embolism. Arch Neurol Psych 81: 667-677 (Jun)
1959
80. Carter AB: Immediate treatment of cerebral infarction. Seven years'
personal experience. Lancet 1: 345-351 (Feb 13) 1960
81. Report of the Working Party on Anticoagulant Therapy in Coronary
Thrombosis to the Medical Research Council: Assessment of short-term
anticoagulant administration after cardiac infarction. Br Med J 1:
335-342 (Feb 8) 1969
82. Veterans Administration Hospital Investigators: Anticoagulants in
acute myocardial infarction. Results of a cooperative clinical trial.
JAMA 225: 724-729 (Aug 13) 1973
83. Harvald B, Hilden T, Lund E: Long-term anticoagulant therapy after
myocardial infarction. Lancet 2: 627-630 (Sep 29) 1962
84. Sibley WA, Morledge JH, Lapham LW: Experimental cerebral infarction: The effect of dicumarol. Am J Med Sci 234: 661-663 (Dec) 1957
85. Moyes PD, Millikan CH, Wakim KG, et al: Influence of anticoagulants on experimental canine cerebral infarcts. Proc Staff Meet
Mayo Clin 32: 124-130 (Mar 20) 1957
85a. Whisnant JP, Millikan CH, Sayre GP, et al: Effect of anticoagulants on
experimental cerebral infarction. Clinical implications. Circulation 20:
56-65 (Jul) 1959
86. Acheson J, Danta G, Hutchinson EC: Controlled trial of dipyridamole
in cerebral vascular disease. Br Med J 1: 614-615 (Mar 8) 1969
87. Acheson J, Hutchinson EC: Controlled trials of clofibrate in cerebral
vascular disease. Atherosclerosis 15: 177-183 (Mar-Apr) 1972
88. Evans G: Effect of platelet-suppressive agents on the incidence of
amaurosis fugax and transient cerebral ischemia. In McDowell FH,
Brennan RW (eds): Cerebral Vascular Diseases. Eighth Conference.
New York, Grune and Stratton Inc, pp 297-299, 1973
89. Gent M: Canadian cooperative study of recent recurrent presumed cerebral emboli (RRPCE). Design and organization of the study. In Hirsh J,
Cade JF, Gallus AS, et al (eds): Platelets, Drugs and Thrombosis. Basel,
Karger, pp 253-257, 1975
90. Blakely JA, Gent M: Platelets, drugs and longevity in a geriatric population. In Hirsh J, Cade JF, Gallus AS, et al (eds): Platelets, Drugs and
Thrombosis. Basel, Karger, pp 284-291, 1975
91. Dyken ML, Kolar OJ, Jones FH: Differences in the occurrence of
carotid transient ischemic attacks associated with antiplatelet aggregation therapy. Stroke 4: 732-736 (Sep-Oct) 1973
92. Harrison MJG, Marshall J, Meadows JC, et al: Effect of aspirin in
amaurosis fugax. Lancet 2: 743-744 (Oct 2) 1971
93. Mundall J, Quintero P, von Kaulla KN, et al: Transient monocular
blindness and increased platelet aggregability treated with aspirin: A
case report. Neurology 22: 280-285 (Mar)'1972
94. Sullivan JM, Harken DE, Gorlin R: Pharmacologic control of thromboembolic complications of cardiac-valve replacement. N Engl J Med
284: 1391-1394, 1971
95 Gent M: Platelet inhibitor therapy in cerebrovascular disease. Thromb
Diath Haemorrh (Suppl) 60: 509-516, 1974
96. Pechet L, Genton E: Fibrinolysis. In Mengel CE, Frei E, Nachman R
(eds): Hematology — Principles and Practice. Chicago, Year Book
Medical Publishers, pp 657-670, 1972
97 Atkin N, Nitzberg S, Dorsey J: Lysis of intracerebral thromboembolism with fibrinolysin. Report of a case. Angiology 15: 436-439
(Oct) 1964
98. Herndon RM, Meyer JS, Johnson JF, et al: Treatment of thrombosis
with fibrinolysin. Preliminary report. Am J Cardiol 6: 540-545 (Aug)
1960
99. Schmutzler R, Koller F: Thrombolytic therapy. In Poller L (ed): Recent
Advances in Blood Coagulation. Boston, Little Brown and Co, pp
299-335, 1969
100. Meyer JS, Herndon RM, Johnson TF, et al: Treatment of cerebrovascular thrombosis with plasmin and plasminogen activators. In Millikan
CH (ed): Research Publications, Association for Research in Nervous
and Mental Disease. Baltimore, Williams and Wilkins, vol 41, pp
373-398, 1966
101. Meyer JS, Gilroy J, Barnhart ME, et al: Therapeutic thrombolysis in
cerebral thromboembolism: Double-blind evaluation of intravenous
plasmin therapy in carotid and middle cerebral arterial occlusion.
Neurology 13: 927-937 (Nov) 1963
102 Meyer JS, Gilroy J, Barnhart ME, et al: Therapeutic thrombolysis in
cerebral thromboembolism: Randomized evaluation of intravenous
streptokinase. In Siekert RG, Whisnant JP (eds): Cerebral Vascular
Diseases. Fourth Conference. New York, Grune and Stratton Inc, pp
200-213, 1965
103. Hass WK, Clauss RH, Goldberg AF, et al: Special problems associated
with surgical and thrombolytic treatment of strokes. Arch Surg 92:
27-31 (Jan) 1966
Fletcher AP, Alkjaersig N, Lewis M, et al: A pilot study of urokinase
104.
therapy in cerebral infarction. Stroke 7: 135-142 (Mar-Apr) 1976
105. Hanaway J, Torack R, Fletcher AP, et al: Intracranial bleeding associated with urokinase therapy for acute ischemic hemispheral stroke.
Stroke 7: 143-146 (Mar-Apr) 1976
Official
S~Qr\
^Association, Inc.
^ «
Journals of the ( J )
^American Heart V 1 /
AMERICAN HEART ASSOCIATION
1976 SUBSCRIPTION RATES
CIRCULATION
Downloaded from http://stroke.ahajournals.org/ by guest on June 18, 2017
A monthly devoted to clinical research and advances in he cardiovascular field.
Eugene A. Stead, Jr., M.D., Editor
U.S.
$25
Elsewhere
$30
CIRCULATION RESEARCH
A monthly concerned with basic research in the cardiovascular field. Brian F.
Hoffman, M.D., Editor
$50
U.S.
$55
Elsewhere
STROKE
A journal of cerebral circulation, bi-monthly. Clark H. Millikan, M.D ., Editor
$25
U.S.
$30
Elsewhere
RECURRING BIBLIOGRAPHY
OF HYPERTENSION
Bi-monthly, in cooperation with the National Library of Medicine.
U.S.
Elsewhere
Publishing Office
AMERICAN HEART ASSOCIATION
7320 Greenville Ave.
Dallas, Texas 75231
$12
$16
XIV. Cerebral ischemia: the role of thrombosis and of antithrombotic therapy. Study group
on antithrombotic therapy.
E Genton, H J Barnett, W S Fields, M Gent and J C Hoak
Stroke. 1977;8:150-175
doi: 10.1161/01.STR.8.1.150
Downloaded from http://stroke.ahajournals.org/ by guest on June 18, 2017
Stroke is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231
Copyright © 1977 American Heart Association, Inc. All rights reserved.
Print ISSN: 0039-2499. Online ISSN: 1524-4628
The online version of this article, along with updated information and services, is located on the
World Wide Web at:
http://stroke.ahajournals.org/content/8/1/150
Permissions: Requests for permissions to reproduce figures, tables, or portions of articles originally published in
Stroke can be obtained via RightsLink, a service of the Copyright Clearance Center, not the Editorial Office.
Once the online version of the published article for which permission is being requested is located, click Request
Permissions in the middle column of the Web page under Services. Further information about this process is
available in the Permissions and Rights Question and Answer document.
Reprints: Information about reprints can be found online at:
http://www.lww.com/reprints
Subscriptions: Information about subscribing to Stroke is online at:
http://stroke.ahajournals.org//subscriptions/